The production and fate of molecular anions formed by electron... compounds

advertisement
The production and fate of molecular anions formed by electron attachment to low electron affinity
compounds
by Michael Jeffrey Salyards
A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of
Philosophy in Chemistry
Montana State University
© Copyright by Michael Jeffrey Salyards (2003)
Abstract:
Resonance electron capture and thermal electron detachment rate constants have been determined for
several low electron affinity (EA) compounds, including anthracene, benzophenone, quinoxaline.
These measurements were taken by comparing the molecule of interest with SF6 using pulsed high
pressure mass spectrometry (PHPMS) to evaluate the time profiles for the relevant anions. These
measurements were affected by re-capture of detached electrons as well as loss of these electrons by
diffusion to the walls of the ion source. This dissertation also explains why these low EA molecules are
not seen at atmospheric conditions. Using the PHPMS, the reactions of the molecular anions of
anthracene, quinazoline, benzophenone, quinoxaline and azulene with oxygen and water have been
studied. In the simultaneous presence of oxygen and water, these molecular anions, M, are rapidly
destroyed and the ion, (O2 H2O), is rapidly formed. The high rate with which this transition occurs
cannot be explained by the simplest model envisioned that is based on well-known ion molecule
reactions. These results can be explained, however, by inclusion into the model of a previously
uncharacterized reaction between the molecular ion-oxygen complex, (M O2)", and water. The results
reported here explain why the molecular anions of compounds that have lower EA’s than that of
azulene are not readily observed in electron capture ion sources of one atmosphere buffer gas pressure.
In addition, it is shown that the reactions characterized here lead to a state of chemical equilibrium
between the M and (O2 H2O) ions within the PHPMS ion source from which the EA values of the
low-EA compounds can be determined. By this method the electron affinities of anthracene,
quinazoline, benzophenone and quinoxaline are found to be 0.54, 0.56, 0.61 and 0.68 eV, respectively. THE PRODUCTION AND FATE OF MOLECULAR ANIONS FORMED BY
ELECTRON ATTACHMENT TO LOW ELECTRON AFFINITY COMPOUNDS
by
Michael Jeffrey Salyards
I
A dissertation submitted in partial fulfillment
of the requirements for the degree
Cf'
Doctor of Philosophy
in
Chemistry
MONTANA STATE UNIVERSITY
Bozeman, Montana
: 'I
June 2003
D r w
APPROVAL
of a dissertation submitted by
Michael Jeffrey Salyards
This dissertation has been read by each member of the dissertation committee
and has been found to be satisfactory regarding content, English usage, format, citations,
bibliographic style, and consistency, and is ready for submission to the College of
Graduate Studies.
Approved for the Department of Chemistry and Biochemistry
Dr Paul A. Grieco
(Gignaturpy^
Date
Approved for the College of Graduate Studies
Dr Bruce R. McLe
(Signature)
Date
iii
STATEMENT OF PERMISSION TO USE
In presenting this dissertation in partial fulfillment of the requirements for a
doctoral degree at Montana State University, I agree that the library shall make it
available to borrowers under rules of the Library. I further agree that copying of this
paper is allowable only for scholarly purposes, consistent with the “fair use” as
prescribed in the US Copyright Law. Request for the extensive copying or reproduction
of this dissertation should be referred to Bell & Howell Information and Learning, 300
North Zeeb Road, Ann Arbor, Michigan 48106, to whom I granted “the exclusive right to
reproduce and distribute my dissertation in and from microform along with the non­
exclusive right to reproduce and distribute my abstract in any format in the whole or in
part.”
Signature (A \
Date
.
“
I
<24 ZTotOt.
2 oc>2>
iv
ACKNOWLEDGEMENT
It is with a heart of thanksgiving that I finish my time at Montana State
University. I cannot imagine how I would have finished or even Started this research
without the tireless efforts of my mentors, Dr Eric Grimsrud and Dr Berk Knighton.
In the immortal words of John Steinbeck’s Seer in Canary Row, “I wouldn’t think
of doing something quite so big without having someone who’s pretty handy with a
sponge in my comer.” My comer is filled with people who love and support me. Debbie,
Amy and Mary - you are the stars in my sky, flooding my life with love, joy, and
support. I am so proud of all you. Mom and Don, Dad and Renee, Rob and Mimi - you
are my heritage. Any success that I have can be traced directly back to each of you. Your
encouragement, love, support, and sacrifice are engraved on my heart. Jason and Jon there is no way to go wrong with brothers like the two of you. Vem and Tanja Brekke,
and all my ukes at Ternion - you have been good friends and dedicated sensies. I will
never forget our journey from white to black. Pastor John Marks and my brothers and
sisters at Abundant Life Fellowship - thanks for your prayers, friendship, and wise
counsel. Finally, the United States Air Force Academy-1 can’t think of a more dedicated
and professional group of people to work with. Thank you for letting me be on your team.
V
TABLE OF CONTENTS
1. INTRODUCTION AND LITERATURE REVIEW...................................................... I
<0 Os UJ is)
Pulsed High Pressure Mass Spectrometry...................................... .
Hardware.......................................................................................
Types of Data Collected................................................................
Plasma Conditions in the Ion Source.............................................
Electron Capture Cross Section, Feshbach States, and M*" Lifetimes..........................13
Interactions of Anions with a Buffer Gas......................................................................24
Stabilizing Collisions and the High Pressure Limit..................................................24
Excitation and Thermal Electron Detachment......................................................... 29
Methods for Measuring Resonance Electron Capture Rates.............................
34
The Suitability of SFe as a Surrogate.............................................................................36
Conclusion...................................................................
37
2. RESONANCE ELECTRON CAPTURE AND THERMAL DETACHMENT............38
Benzophenone...............................................................................................................38
Experimental.............. :..................................................................... :..................... 39
Results and Discussion..............................................
39
Apparent Loss of BP" to SFe.............................................................................. 41
Time Profile Data and Differential Equations Model........................................ 43
Failure of Model at Low Number Densities....................................................... 48
The New Model with Electron Diffusion............................................................51
Predictions of Thermal Detachment....................................................................53
Efforts to Stabilize BP" with Sip4............................................................................ 54
Quinoxaline...................................................................................................................55
Experimental......................................
55
Results and Discussion.............................................................................................57
Anthracene.................................................................................................................... 60
Experimental.............................................................................................................61
Results and Discussion.............................................................................................62
Quinazoline................................................................................................................... 66
Experimental...................................................................................................
67
Results and Discussion.............................................
68
Conclusion.................................................................................................................... 72
vi
TABLE OF CONTENTS - CONTINUED
3. FAILURE TO DETECT THESE ANIONS AT ATMOSPHERIC CONDITIONS..74
Introduction..................................................................................................
74
Experimental..................................................................................................................78
Results and Discussion.................................................
79
Determination of Mechanism........................................................
81
Determination of Electron Affinities........................................................................94
Conclusions...................................................................................................................98
REFERENCES CITED.....................................................................................................99
APPENDICES................................................................................................................I l l
APPENDIX A: HIGH PRESSURE LIMIT EQUATIONS.................... .......... ,........112
APPENDIX B: REVffiW OF PSUEDO 1st ORDER KINETICS.............................. 116
APPENDIX C: MODELING WITH DIFFERENTIAL EQUATIONS......................118
vii
LIST OF TABLES
Table
Page
1-1. Review of Feshbach resonance descriptions in the literature.................................. 20
1-2. Thermodynamic and kinetic properties of a typical molecule................................. 33
2- 1. Parameters used for the IBM Kinetic Simulator.................................................... 45
2-2. Thermodynamic and kinetic properties of benzophenone....................................... 54
2-3. Thermodynamic and kinetic properties of quinoxaline.................. .......................... 59
2-4. Thermodynamic and kinetic properties of anthracene......................... .................... 66
2-5. Summary of kinetic and thermodynamic properties for each molecule....................72
viii
LIST OF FIGURES
Figure
Page
1-1. Schematic of the PHPMS.............................. ............................................................. 3
1-2. Photograph of the PHPMS front flange.......................................................................4
1-3. Standard and logarithmic time profiles of the SFe anion.............................................6
1-4. Logarithmic and normalized time profiles of two anions............................................8
1-5. Normalized time profiles showing equilibrium of several ions...................................9
1-6. Nitric oxide data showing the ion dominated, plasma transition ...........
12
1-7. Fluorocarbon ion lifetimes versus vibrational degrees of freedom...................... .....22
1-8. Unimolecular rate constant versus pressure for a theoretical case........... .................26
1- 9. Comparison of thermal electron detachment predictions.............
33
2- 1. Structure of benzophenone......................................................................................38
2-2. Mass spectra of benzophenone and SFe at different temperatures..... .......................40
2-3. Time profiles of the ions in Figure 2-2 ....................................
........41
2-4. Measured LtranSfervalues for electron transfer to SFe................................................. 43
2-5. Time profiles of BP' and SFe" at different temperatures........................
44
2-6. Extrapolation of initial concentrations for logarithmic time profiles........................ 47
2-7. Measured electron capture rate constants for benzophenone.................................... 47
2-8. The linear model of Lec and Laet for benzophenone........,..........
50
2-9. Histogram of calculated electron diffusion rate constants...........
52
2-10. Predicted versus measured BP7slopes from logarithmic time profiles...................53
2-11. Prediction of thermal electron detachment rate constants for benzophenone......,..54
ix
LIST OF FIGURES continued
2-12. Mass spectra of (BPoSiF4)" and SFg"....................................................................... 56
2-13. Structure of quinoxaline......................................................... ............................... 57
2-14. Measured electron capture rate constants for quinoxaline...................................... 58
2-15. Predicted versus measured Qx slopes from logarithmic time profiles................... 59
2-16. Prediction of thermal electron detachment rate constants for quinoxaline.............60
2-17. Structure of anthracene.................................................................................:......... 61
2-18. Effects of 179? amu impurity on anthracene........... .............................................. 62
2-19. Measured electron capture rate constants for anthracene........................................63
2-20. Measured electron capture rate constants for anthracene.........................................64
2-21. Predicted versus measured A" Slopes from logarithmic time profiles.....................65
2-22. Prediction of thermal electron detachment rate constants for anthracene...............67
2-23. Structure of quinazoline.......................... ............................................................... 67
2-24. Mass spectrum of Qz" quinazoline and O2 cluster..................................................69
2-25. Measured electron capture rate constants for quinazoline...................................... 70
2- 26. Loss of Qz to thermal detachment....................................................................... 71
3- 1. IMS spectra of anthracene and benzophenone........................................................77
3-2. Negative ion mass spectra of oxygen and water and anthracene...............................80
3-3. Time profile of the family of anthracene, oxygen and wateranions.......... ................ 82
3-4. Computer generated time profiles of the Model A mechanism................................. 85
3-5. Computer generated time profiles of the Model B mechanism.................................87
3-6. Time profile of the family of quinazoline, oxygen and water anions
88
X
LIST OF FIGURES continued
3-7. Time profile of the family of benzophenone, oxygen and water anions....................90
3-8. Time profile of the family of quinoxaline, oxygen and water anions .......................91
3-9. Time profile of the family of azulene, oxygen and water anions...............................93
3-10. Electron affinity determinations for each molecule in this study,
96
xi
ABSTRACT
Resonance electron capture and thermal electron detachment rate constants have been
determined for several low electron affinity (BA) compounds, including anthracene,
benzophenone, quinoxahne. These measurements were taken by comparing the molecule
of interest with SFg using pulsed high pressure mass spectrometry (PHPMS) to evaluate
the time profiles for the relevant anions. These measurements were affected by re-capture
of detached electrons as well as loss of these electrons by diffusion to the walls of the ion
source. This dissertation also explains why these low BA molecules are not seen at
atmospheric conditions. Using the PHPMS, the reactions of the molecular anions of
anthracene, quinazoline, benzophenone, quinoxaline and azulene with oxygen and water
have been studied. In the simultaneous presence of oxygen and water, these molecular
anions, M ", are rapidly destroyed and the ion, (O2 0HaO)", is rapidly formed. The high
rate with which this transition occurs cannot be explained by the simplest model
envisioned that is based on well-known ion molecule reactions. These results can be
explained, however, by inclusion into the model of a previously uncharacterized reaction
between the molecular ion-oxygen complex, (M0O2)", and water. The results reported
here explain why the molecular anions of compounds that have lower BA’s than that of
azulene are not readily observed in electron capture ion sources of one atmosphere buffer
gas pressure. In addition, it is shown that the reactions characterized here lead to a state
of chemical equilibrium between the M" and (O2 0HzO)" ions within the PHPMS ion
source from which the BA values of the Iow-EA compounds can be determined. By this
method the electron affinities of anthracene, quinazoline, benzophenone and quinoxaline
are found to be 0.54, 0.56, 0.61 and 0.68 eV, respectively.
I
INTRODUCTION AND LITERATURE REVIEW
One of the most fundamental processes in chemistry, especially analytical
chemistry, is the capture of an electron by a neutral molecule to form the parent anion.
This process, often referred to as resonance electron capture (REC), can be naively
written as Reaction 1-1:
kec
M + e~ ^ M"
(i-i)
kdet
where Icec is the electron capture rate constant, and Icdet is the thermal electron detachment
rate constant. Some of the sections that follow will show that this reaction also involves
an intermediate. The importance of electron capture and negative ion formation to
analytical chemistry has its roots in the 1948 discovery of a young English scientist. At
that time, James Lovelock was commissioned by the British Medical Research Council to
study the common cold. Specifically, he was asked to investigate the popularly held
notion that colds are caught by being exposed to a cold draft. Lovelock’s efforts, which
included using radioactive a particle emitting paint from WWII aircraft gauges, to
measure changes in the density of air led to the invention of an ionization apparatus. He
did not learn anything useful about the common cold, but he did succeed in making a
device that was especially sensitive to halocarbon gases [I]. Eventually the ionization
source was changed to a p emitter, and the electron capture detector (ECD) was bom.
ECD use in the next 10 years moved to biochemistry and then to the detection of
environmentally dangerous molecules [2] '.
In 1962, Rachel Carson published Silent Spring [3]. Writing somewhat poetically
about the environment, she described an impending chemical apocalypse. The ECD had
proven so sensitive that researchers reported finding trace levels of pesticides in samples
ranging from penguin fat to mother’s milk. Carson’s work thrust the ECD into the
scientific limelight, and the study of electron capture and the formation of gaseous ions
became popular. Presently, the study of electron capture is important in a wide range of
disciplines including gaseous dielectrics [4], excimer lasers [5], discharges used for
etching [6], and atmospheric chemistry [7].
The sections that follow will provide a discussion and review of: a) the pulsed
high pressure mass spectrometer (PHPMS) which was used to collect data for all of the
experiments in this dissertation, and b) relevant theories and literature surrounding the
phenomenon of electron capture and its opposite, thermal electron detachment. Chapter 2
describes experiments that were designed to measure resonance electron capture and
thermal electron detachment rate constants for several low electron affinity (BA)
molecules. Finally, Chapter 3 presents a series of experiments that explain why these low
EA molecules are difficult to detect at atmospheric conditions.
Pulsed High Pressure Mass Spectrometry
Kebarle developed the PHPMS during the late 1960’s, and its construction and
operation have been fully described previously [8]. This section is not intended to repeat
the details of construction provided in this reference. However, brief descriptions of the
3
physical apparatus, the ionization source, the electron optics, and the types o f data that
can be collected, may be helpful for understanding the experiments in Chapters 2 and 3.
Hardware
Figure 1-1 is a schematic o f the PHPM S. The gas handling plant is a 6.5 liter glass
bulb. It is typically pressurized to approxim ately 800 torr with a buffer gas.
Gas Handling Plant
Quadrupole
Mass Spectrometer
W ISource
Electron Optics
and Pulsing
Heating
Filament
Figure 1-1. A schematic representation o f the PHPM S hardware and main components.
For all o f the experim ents in this dissertation, m ethane was used as the buffer gas. Other
gases including helium, nitrogen, oxygen, and argon with 15% m ethane have been used
in previous experim ents [9]. The gas handling plant is fitted with an injection port that
allows introduction o f both gas and liquid samples. Solid samples are dissolved in a
suitable solvent (toluene was often used for the experiments in this dissertation) and the
solution is injected. The gas handling plant is typically heated to about 120°C to ensure
4
com plete volatilization o f sample m olecules that often have relatively low vapor
pressures.
Once the gas handling plant is loaded with the desired analytes, a valve is opened,
and the m ixture flows through a heated transfer line into the ion source as shown in
Figure 1-2.
Figure 1-2. A photograph o f the front flange o f the PHPMS revealing the transfer line,
ion source and the electron filam ent and pulsing optics. In this orientation, the gas
mixture would flow from the bottom o f the picture frame to the top. The electron optics
are 90° off axis to the right at the top o f the picture. For reference, the flange
circum ference is 10 inches.
The pressure in the ion source is controlled by the am ount of gas flowing out of the gas
handling plant. Typical ion source pressures are between 1-5 torr, and the ion source is
heated between 30-150°C. Both o f these param eters are adjusted as dictated by a given
experiment.
5
Electrons are produced from a heated filament that is -2.95 kV relative to the ion
source. A gate lens is approximately 0.75 cm in front of the filament. The gate lens is
held at -3.0 kV. Approximately every 75-100 msec the gate lens potential is pulsed to
- 2.90 kV for about 20 jisec. This voltage change sends a pulse of 2.95 keV high energy
electrons through a stack of focusing lenses into the ion source via a small entrance
aperture. Because the number density of the buffer gas is about 7-9 orders of magnitude
greater than the number density of the sample molecules, these high energy electrons are
believed to interact primarily with the buffer gas. The chemical cascade that follows was
first described by Munson and Field in their seminal paper on chemical ionization [10].
Basically, a whole host of positive methane and ethane related ions are formed as well as
a complementary population of thermal electrons. In methane buffer gas, each high
energy electron produces about 85 thermal electrons. If the sample molecules have a
positive BA, they will capture some of these thermal electrons and form anions.
Depending on their proton affinity, sample molecules may also form cations by
interacting with the positively charged buffer gas related ions. As will be detailed in the
next section, these newly formed ions diffuse to the wall of the ion source. A
representative population of these ions passes through an exit slit into a quadrupole mass
spectrometer. By narrowing the exit slit, the ion source can be operated in the 10-20 Torr
region. However, the narrow slit decreases the ion signal. Ions are detected using an '
electron multiplier and an ion counting channeltron.
6
Types o f D ata Collected
The electron pulsing and related detection hardware and softw are are what make
the PHPM S unique. M ultichannel scalar software allows discrete channels o f data to be
collected after the electron pulse event. The time length o f the channels can be varied
from 5 fisec to several hundred msec. By selecting long channel tim es (=160 msec) and
scanning the quadrupole, typical mass spectra are obtained. H owever if much shorter
channel times (= 100 (isec) are used and the appropriate m/z is selected on the
quadrupole, the population o f a given ion in the source can be tracked in time as it is
formed and diffuses to the wall. As might be expected, the diffusion o f ions to the wall is
a first order loss process and indicated in Figure 1-3.
6000 -I
5000 R- 4000
m 3000
.g> 2000
Time (msec)
Time (msec)
Figure 1-3. Standard time profile o f SFe anion. The ion source pressure is = 3.0 Torr. The
ion source tem perature is 60°C. (A) is the standard plot of signal versus time. (B) is the
same data plotted as the In o f the signal versus time. A line as been superim posed on the
data to highlight the linear relationship.
I
The diffusion rate constant, kdiff, is approximately 150 sec"1 at 3 torr and 60°C. Diffusion
of ions depends primarily on three variables, the temperature squared, the inverse of
pressure, and the reduced mobility of the ion. Equation 1-2 can be derived from equations
provided by McDaniel and Munson [11].
iji2
^diff “
(1-2)
In this equation, the reduced mobility term, Ko, contains information about the reduced
mass of the ion of interest as well as a term that describes how strongly the ion interacts
with the buffer gas [12]. For all the experiments in this dissertation, ions were considered
to have the same mobility to a first approximation. Therefore, diffusion is treated as a
function of inverse pressure and temperature squared.
In addition to diffusion, ions can be lost if they react with a neutral molecule or
another ion. Further, the diffusional loss of other ions may be partially offset if they are
being produced by a chemical reaction. Sometimes it is helpful to compare the time
profiles of two ions as in Figure 1-4. Close examination, of the logarithmic plot in Figure
1-4 (A) reveals that the population of the ion represented by the circles is disappearing
faster than the population of the ion represented by the squares. Because these ions are
both being lost to diffusion, sometimes these other loss and production terms are subtle
and difficult to see on a logarithmic plot. The normalized plot in Figure 1-4 (B) is
prepared by dividing a given ion signal by the total ion signal. The inference from Figure
1-4 (B) is that the circle ions are reacting away and the square ions are being produced.
8
■0.80 M 0.60 0.40 0.20
Time (m sec)
-
Time (msec)
Figure 1-4. (A) A logarithm ic plot similar to Figure 1-3 (B) but o f tw o ions, (o), and ( ).
(B) The same data shown as a norm alized plot. The normalized plot demonstrates how
each ion population is changing relative to the total population. In effect, this type of
graph removes diffusion and (assum ing ions are not diffusing at different rates)
em phasizes any ion molecule chem istry that may be occurring.
Norm alized time profiles may also reveal if two or more ions have com e into
equilibrium as shown in Figure 1-5. During the first 5 msec of this experim ent, the ions
represented by (x) appear to be reacting aw ay while the ions represented by (o) and (A)
are being produced. After about 5 msec, the relative concentrations becom e constant in
time. This type o f profile is strongly suggestive o f chem ical equilibrium . However, it
should be noted that this steady state condition w ould also be observed for two ions with
sim ilar diffusion coefficients that are not reacting with each other at all. As shown in
detail in Chapter 3, verifying equilibrium involves varying concentrations o f reactants
and products and com paring observed equilibrium constants with expected values.
9
Time (msec)
Figure 1-5. Norm alized time profiles o f several different ions showing w hat equilibrium
typically looks like.
Plasm a Conditions in the Ion Source
After the initial pulse o f high energy, 2.95 keV, electrons enter the source, a weak
plasm a is created. Initially, the positive charge carriers are prim arily ions materially
related to the methane buffer gas (i.e. CHs+, and C 2 Hs+) and, the negative charge carries
are thermal electrons. Several events happen rapidly. First, anions can be formed if some
o f the electrons are captured by sample m olecules with a reasonable EA and a high
enough electron capture rate constant. None o f the experiments in this dissertation deal
directly with positive ion chemistry, but it should be noted that the buffer gas related
cations will also charge transfer to molecules with suitable proton affinities. The resulting
plasm a contains positive ions, negative ions and thermal electrons.
10
Both Kregel and McDaniel have eloquently described how this type of plasma
behaves as the charged species diffuse to the wall of the ion source [13, 14]. Most
recently, Franklin provided an outstanding review of electronegative plasmas [15]. A
qualitative review of,these works follows. Initially, the electrons diffuse much faster than
the ions due to their high mobility. Biondi reports the mobility of an electron to be 1000
times that of a negative ion [16]; however, Franklin states the mobility of an electron to
be 300 times that of a negative ion [15]. Neither author provides data or theory to support
their reported electron mobility. However, it is generally accepted that mobility of an
electron is high compared to an anion. This high mobility is the result of the electron’s
mass (it is =275,000 times smaller than a 150 amu anion) Offset by inelastic collisions
and interactions with the buffer gas. Electrons and all charged particles are neutralized at
the wall. So, a charge imbalance develops. The plasma by definition maintains charge
neutrality. As a result the diffusion of positive ions becomes coupled to the diffusion of
electrons. The net effect is that the diffusion rate of the positive ions is doubled and the
diffusion rate of the electrons is retarded to match the positive ions. This condition is
often called ambipolar diffusion, and this type of plasma is referred to as electron
dominated or an electron/positive ion plasma. The negative ions are, as Kregel states,
“bottled” in the ion source and their diffusion rate is = 0. This electron dominated plasma
persists until the electron density is < 10% of the negative ion density. As the electrons
are depleted, the nature of the plasma changes to a negative ion dominated or negative
ion/positive ion plasma, and the negative and positive ions begin to diffuse at their
11
normal rates. Miller provides a good mathematical summary of the ambipolar diffusion in
an electron dominated plasma, listed here as equations 1-3,1-4 and 1-5 [17]:
(1-3)
(1-4)
n
D
(1-5)
e'
e
where Daxis the ambipolar diffusion coefficient of x, and Dx is the free diffusion
coefficient of x (i.e. in the absence of a space charge), and nx is the number density of x.
Note that Equations 1-3 - 1-5 are in terms of a diffusion coefficient, D, instead of a
diffusion rate constant, kdiff. These terms are linearly related as in Equation 1-6:
D
( 1- 6)
where A is the characteristic diffusion length. If the ion source is a sphere, A = the radius
of the sphere divided by K. McDaniel provides more details about how to calculate A for
ion sources with atypical geometries [11, 14],
The ambipolar diffusion coefficient of negative ions, Da,., is ~ 0 because (D. ZDe)
~ 0.001 - 0.003 (the mobility of an electron is 300 - 1000 times greater than the mobility
of an ion as previously mentioned). For a more rigorous treatment that describes both
12
electron dom inated and ion dom inated am bipolar diffusion see R ogoff [18]. Lineberger
and Puckett reported this type o f behavior in a stationary after glow instrum ent (sim ilar to
the PHPM S) in 1969 [19]. They used nitric oxide to make a plasm a that contained
(N O N O )+, N O 2" and free electrons as the charge carriers. Figure 1-6 is a recreation o f
their data.
At about 150 msec, a transition occurs. Before this point, the positive ions are
diffusing approxim ately tw ice as fast as they are after the transition. The negative ions
are not diffusing at all before the transition. After the transition both the positive and
c § 0.6
U
S 0 .4
Time (msec)
Figure 1-6. A recreation o f Lineberger’s nitric oxide data. The positive ions are
represented by the dotted line; the negative ions are represented by the solid line.
Electrons are not shown directly in this graph, but their effect is clearly seen at 150 msec
where the plasm a changes from electron dom inated to ion dominated. The positive ion
signal has been translated up to make it easier to see. After the transition point it should
lay on top o f the negative ion signal.
13
negative ions are diffusing at the same rate. Reviewing the logarithmic plots in Figures 13 (B), and 1-4 (A) shows that during the first few msec these anions are “bottled” inside
the source.
When the number of charged particles becomes sufficiently small, the plasma
condition breaks down, and the diffusion rates of electrons, negative ions, and positive
ions are no longer coupled. This condition is often referred to as an ionized gas, and
diffusion coefficients are referred to as free to distinguish them from ambipolar.
Data collected from the PHPMS for experiments in this dissertation are taken in
the ion dominated ambipolar diffusion and the ionized gas free diffusion regions. If the
behavior of two ions is being compared, it is not critical that the experiment be carefully
limited to one type of diffusion as long as both ions are being affected equally by changes
in the diffusion. As will be seen when thermal electron detachment is discussed in
Chapter 2, the transition from an ion dominated plasma to free diffusion can have a huge
effect on the recapture of detached electrons.
Electron Capture Cross Section, Feshbach States, and M*' Lifetimes
The term resonance electron capture rate is rather unfortunate for two reasons.
First, it is also used to describe the completely unrelated nuclear reaction where an inner
shell electron is captured by the nucleus as a proton is transformed into a neutron.
Second, the term has a connotation that is not obvious. It is often used by gas phase ion
chemists to describe electron capture reactions^ like Reaction 1-1, where a molecular
anion is formed. In contrast, the term dissociative electron capture is used to describe an
electron capture event where a fragment ion is formed.
The concept of resonance electron capture is deceptively complex. Electron
capture rates are a seemingly fundamental property. And, one might expect a molecule
with a positive EA to capture electrons at some collisional rate analogous to Langevin or
ADO theory in ion molecule chemistry. Even if electron capture rates differ for each
molecule, shouldn’t the typical junior level physical chemistry text have catalogue of
electron capture rate constants for common molecules? Unfortunately, measuring
electron capture rates is not easy. The theory is complicated and involves a failure of the
Bom-Oppenheimer approximation. Further, as shown by Knighton et al [20], electron
capture rates do not seem to be a function of any standard thermodynamic properties. The
first step to understanding resonance electron capture involves moving beyond Reaction
1-1 and writing the more complete Reaction 1-7:
ki
M + e' ^
k_i
k2[B]
M
(i-7)
k_2[B]
This section will focus on the formation and lifetime of the excited intermediate, M*" or
the first half of this reaction. The following section will focus on the interaction of the
excited intermediate with the buffer gas, or the second half of the reaction.
The initial formation of a temporary or transient negative ion (TNI) is most often
described in the literature in terms of an electron capture cross section, Crcap. This cross
section, first given by Vogt and Wannier [21], is Equation 1-8:
15
Ocap = 4 t o q -Ja/2E
(1 . 8 )
where al is the Bohr radius, and a and E are the polarizabilty and the electron energy
respectively (in atomic units). This equation is thought to be valid for very low electron
energies, E ~ 0. As electron energy increases, the de Broglie s-wave cross section (given
in Equation 1-9) may be more appropriate [22]:
G
Tra0
— ------cap
2E
(1-9)
Equation 1-10, from Klots ’ highly referenced 1976 paper [23], is an analytical form of
this equation that describes the cross section in terms of the electron energy:
i
G cap
= TtX2H - e1'4^
1)
(M O )
where X is the de Broglie wavelength of relative motion. The Langevin parameter, y, is
given by equation 1-11:
Ry 2
( 1- 11)
where |i is the mass, m is the mass of an electron, R is the Rydberg constant, a the
polarizability, and a0 the Bohr radius. Equation 1-12, provided by Dunning, is a more
convenient version of Equation 1-10:
G cap
Tra^
/6 a 0
/I
2 E (l
-
J-4(2aE)2 ]
V
''
( 1- 12)
16
The electron capture cross section is believed to be related to ki from Reaction
1-7 by Equation 1-13:
CO
(1-13)
where f(v) is the electron velocity distribution as given by Compton et al [24] in
Equation 1-14:
(1-14)
Perhaps the most simple and intuitive form of Equation 1-13 comes from Hahndorf and
Illenberger [25] in Equation 1-15:
where the right hand side of the equation is given in terms of the product of the average
cross section and the average electron velocity for a given temperature.
Because the cross section term decreases with electron velocity, these terms off
set each other, and electron capture rates are believed to be insensitive to temperature
over a large range, 25-145°C in one study [25] and 25-325°C in another [26].
Once the electron is initially captured, an excited M*" is formed. If the neutral
molecule has a positive EA, the M*" species has an excess energy which is equal or
greater to the EA. The lifetime of this intermediate is described by Equation 1-16:
17
I
T = ~—
d-16)
where k _i is the first order autodetachment rate constant from Reaction 1-7.
This lifetime appears to depend on how the electron is bound and where the
excess energy is stored. Different authors describe M*' with different levels of
complexity. In the most simple model, the M*" is considered to be one of two types, short
lived or long lived. The so called long lived species is often termed a Feshbach resonance
[27]. Christophorou describes four different negative ion resonant states for the M*~
species: a) shape or single particle resonances, b) core excited shape resonances, c)
nuclear excited Feshbach resonances, and d) electron excited Feshbach resonances [4].
Both types of shape resonances result when the incident electron is trapped by the
centrifugal motion that arises from the relative motion of the electron and the neutral
molecule. The trapped electron, in this case, will have a certain angular momentum (i.e. 5
> I, 2,). As a result, this channel for anion formation is sometimes called p-wave
attachment for 5 =1. Barsotti et al recently (2002) published a paper showing evidence
of p-wave electron attachment to CI2 [28]. The core excited shape resonance involves the
same concept but begins with the molecule in an excited electronic state.
The nuclear excited Feshbach resonance involves coupling of the kinetic energy
of the electron with rotational and vibrational motion of the mqlecule. This obvious
18
violation of the Bom-Oppenheimer approximation makes actual calculations of the
potential energy surfaces not possible with currently accepted theories.
The electron excited Feshbach resonance happens when the incoming electron
electronically excites the molecule as it is captured. The electronically excited, M*",
offers less shielding of the nucleus, and the incoming electron experiences a slightly
positive charge that can temporarily bind it to the nucleus.
Another way of discussing electron molecule interactions is in terms of the
lifetime of M*". In unpublished lecture notes. Reddish [29] describes three different
scenarios in terms of the lifetime of M*", versus a typical vibrational period, ~ IG"14 sec:
a) impulse limit where T « IO"14 sec, b) shape resonance where T ~ IO"14, and c)
Feshbach resonance where T » IO'14.
The impulse limit applies to the case where the electron is almost elastically
scattered. The negative ion decays before a complete vibration occurs; the molecule can
be rotationally and or vibrationally excited depending on how much energy the electron
lost during the interaction.
Reddish describes the shape resonance similar to Christophorou (above) noting
that the linear momentum of the electron is transferred into angular momentum. Reddish
adds that the incoming electron attaches to the ground state molecule by entering the
lowest unoccupied molecular orbital. Reddish does not address the electron excited shape
resonance.
The Feshbach resonance, according to Reddish, is similar to the electron excited
Feshbach resonance as described by Christophorou. The incident electron electronically
19
excites the molecule and then becomes bound to the molecule. The result is a negative
ion with two excited electrons and an inner hole. The decay of this ion is believed to be a
concerted process where both electrons have to change states simultaneously (one is
ejected while the other returns to the ground state.) Reddish does not address the nuclear
excited Feshbach resonance.
Most researchers agree that long lived M*" Feshbach resonance species exist
because most detection schemes require the M*" species to live long enough to be
detected or (as will be presented in the next section) Stabilized by collisions with a buffer
gas. But, how researchers describe the Feshbach resonance varies between the nuclear
excited, and the electron excited definitions provided above. Thoss and Domcke have put
forth a theory about vibrational relaxation that is similar to Christophorou’s nuclear
excited Feshbach resonance with one complexity as shown in Reaction 1-17:
ki
krvR
kztB]
M + e~ ^ M**" ^ M*"^ M"
k-i
kpvE
(1-17)
k_2[B]
where IVR refers to intramolecular vibrational relaxation, and IVE refers to
intramolecular vibrational excitation. In this model, the incoming electron vibrationally
excites the molecule. But, (possibly based on symmetry selection rules) only a few
vibrational modes can be excited by the electron. These active modes are coupled to all of
the other vibrational modes. As a result, vibrational energy that was initially concentrated
in the active modes can be redistributed among the other inactive modes. In the parlance
20
of Reaction 1-17, M
relaxes into M*". In order for the excited anion to autodetach, the
energy has to be collected back into the active modes, a process that is not entropically
favored. Many researchers referenced by Thoss and Domcke [30] adhere to this
intramolecular vibrational relaxation model. Table 1-1 summarizes how various
researchers have defined the excited state of the Feshbach resonance (electronic,
vibrational, combination, or undetermined).
Table 1-1. Review of Feshbach resonance descriptions in the literature
Researcher
Molecules Studied
Feshbach Resonance Description
Dessent et al [31]
Nitromethane halide clusters
Vibrational
Leber et al [32, 33]
Nitrous oxide clusters and
carbon dioxide
Vibrational for both and but
mentions electronic for (N 20 ) n
Tulej et al [34]
Porpadienylidene anion
Electronic
M illeretal [35, 36]
SF5CF3, PCl3, and POCl3
Undetermined
Miller et al [37]
SF4 and SF5
Electronic
Hahndorf and
Dlenberger [25]
CF3I, CF3Cl, CF2Cl2,
C6F5Cl, and C6F5CN
Vibrational
Suess et al [38]
SF6, C6F6, CioFg, and C-C7F14 Vibrational
Michaud et al [39]
Oxygen
Combination
Compton et al [24]
SF6, and C6H5NO2
Vibrational
Naff et al [40]
Alicyclic and aromatic
fluorcarbons
Vibrational
Klots [41]
Theoretical treatment
Vibrational
21
Table 1-1 continued.
Researcher
Molecules Studied
Feshbach Resonance Description
Hadjiantoniou et al
[42,43]
Benzene derivatives, various
other organic molecules
Vibrational
Collins et al [44]
p-benzoquinone
Combination,
Johnson et al [45]
NO2 benzene derivatives
Vibrational
Horacek et al [46]
Hydrogen iodide
Vibrational
Garrett [47]
Theoretical treatment
Undetermined
Tobita et al [48]
Polycylic aromatic
hydrocarbons
Combination
Thoss and Domcke
[30]
Theoretical treatment
Vibrational
The most compelling evidence for the vibrational Feshbach resonances comes from data
collected by Naff et al [40] shown here as Figure 1-7. However, Johnson et al directly
referenced this data and pointed out:
x is affected by a number of other factors besides N, such as the
internal energy of the ion (itself a function of electron affinity,
incident electron energy, and the thermal and zero-point energies),
geometry, and configurational changes which may take place
upon capture. If for a group of molecules these other factors are
approximately the same, then it is possible to see an increase in T
with increasing N. [45].
In this quotation, N refers to the number of vibrational degrees of freedom for a given
molecule.
22
io o o O
$
OT
5
d>
O
100 -
S
§
O
O
O
10 -
I-I
I -
10
I
I
I
I
I
20
30
40
50
60
Vibrational Degrees o f Freedom
Figure 1-7. A recreation o f data from N aff and Cooper on alicyclic and aromatic
fluorocarbons, showing the increasing lifetime o f tem porary negative ions from 7
different molecules as a result o f increasing vibrational degrees o f freedom.
Tulej et al provides spectroscopic data o f the popadienylidene anion that supports the
electronic Feshbach resonance [34], and M iller et al suggest that SF^*" is formed by an
electronic Feshbach resonance [37].
Garret (and references within) [47] and Dessent et al [31] discuss the need for the
neutral m olecule to have a dipole m om ent above a critical value, (Ic = 2 debye, in order to
initially bind the electron. The dipole bound electron orbital is believed to be very diffuse
(> 30 Angstroms). The lingering dipole bound electron is then thought to be captured into
the lowest unoccupied m olecular orbital to form the vibrational Feshbach resonance.
As will be seen in Chapter 2, the m olecules in the present study are
benzophenone, anthracene, quinoxaline, and quinazoline. The experim ents in this
dissertation were not designed to determ ine what type o f Feshbach resonances these
23
molecules form. However, based on previous experiments on similar molecules [42, 43,
49], it is very likely that the molecules in the present study form vibrationally excited
Feshbach resonances.
The magnitude of ki and Li cannot be predicted for any molecule, but Compton et
K from first principles as indicated in Equation
al proposed a way to predict the ratio -—
k-i
1-18 [24]:
fc,
p
Z ic V fE A + a f r f - 1
k r p ° ~ M^v(N-I)J flh v ,
( 1 " 1 8 )
z-7
where p" is the density of states for the anion, and p0 is the density of states for the
neutral plus the electron. The variables on the far right hand side of the equation are as
follows: m is the mass of the electron; v is the velocity of the electron; N the number of
vibrational degrees of freedom; EA is the electron affinity; pz is the zero point energy; Vj
the ith vibrational frequency of the anion, and the variable a is an empirical correction
factor that would need to be evaluated for each molecule. Knighton et al provides a
helpful review of Equation 1-18. and the previous attempts to use known -— values to
k -i
estimate EA values (note that Equation 1-18 has a typographical error in the Knighton
paper - the II term in the denominator has been replaced by a E term) [20].
24
Interactions of Anions with a Buffer Gas ,
The ideas in this section should be intuitive to even the most casual observer, for
we have all experienced ordinary gases and liquids coming into thermal equilibrium.
Every day research is conducted in this regard as cold milk is poured into hot coffee and
air conditioners are adjusted to make a room more comfortable.
The intermediate, M*", formed in Reaction 1-7 faces two different outcomes. It
can autodetach the electron as governed by k _i, or it can be stabilized by collisions with
the buffer gas as governed by k2[B], where B is the concentration or pressure of the
buffer gas. Further, the stable MT can be excited back to M*" by collisions with the buffer
gas as governed by k_3[B] , Both the forward and reverse step will be explained here in
turn.
Stabilizing Collisions and the,High Pressure Limit
The Lindemann-Hinshelwood mechanism dating back to 1921 is the starting
point in most text books for unimolecular reactions [50]. These reactions are often written
as Reaction 1-19:
ka
kb
A + B ^ A* ^
k -a[B]
Products
(1-19)
Where B is a buffer gas molecule. Note that Reaction 1-19 is essentially the reverse of the
last part of Reaction 1-7. By making a steady state assumption for the [A*], the reaction
rate can be written as Equation 1-20.
25
d[P roducts]
dt
r A i
k ak b [ B ]
d " 20)
= [ A 1 k J B l + kb
This equation is often given in terms of a unimolecular rate constant as in Equation 1-21:
d [ P r o d u c ts ]
dt
K J A 1;
K K I
b
I-
K J B J + K
( 1- 21 )
The result of the Lindemann-Hinshelwood mechanism is a reaction that behaves
one way at low [B] and another way at high [B]. This behavior becomes intuitive by
evaluating the two extremes: a) the limit of kuni as [B] goes to zero, and the limit of kuni as
[B] goes to infinity. The former case, sometimes called the low pressure limit, is shown
here as Equation 1-22:
^uni
- k a[ B ]
l i m[ B] ^Q
(1"22)
and the latter case, sometimes called the high pressure limit, is shown here as Equation
1-23.
lim[ B 7-^co
(1-23)
26
Note that in the low pressure lim it the reaction behaves like a second order process (i.e.
the reaction rate depends on both [A] and [B]). H owever in the high pressure limit, the
reaction behaves like a first order process (i.e. the reaction rate depends only on [A]).
Graphically, Reaction 1-19 would behave as indicated by Figure 1-8.
2000 i
1600 -
1200
-
400 -
0.000
0.004
0.006
Pressure B (ton)
0.008
Figure 1-8. A representation o f the Lindem ann-H inshelw ood m echanism showing both
the high and low pressure regions. The following parameters were used for this figure:
ka = 2.0 x 10 10 cc sec"1, k_a, 2 x IO'9 cc sec"1, and ky = 1.8 x IO4 sec"1.
As a historical aside in the early 1920’s, Professor J. Perrin exam ined data from a
unim olecular dissociation reaction that, unbeknow nst to him, was in the high pressure
limit. Clearly, the reaction rate seemed independent o f [B], and Perrin extrapolated these
results to very low pressures, and he began to ponder. If the reaction proceeds at the same
rate regardless o f [B], and perhaps in the absence o f [B], what outside agency is causing
the reaction to happen? Perrin soon after put forth his theory that electrom agnetic
27
radiation was responsible for creating the excited intermediate. In 1922, a discussion was
published in which a remarkable group of scientists, Lindemann, Arrhenius, Langmuir
and Dhar, rebutted this theory [51]. Lindemann was especially articulate pointing out that
the unimolecular inversion of sucrose proceeds at the same rate in the sunlight as it does
in total darkness. He also pointed out that the incident sunlight probably only penetrates
the first Imm of the solution. Lindemann then remarked, “After all this criticism, I
suppose I had better put forward a constructive suggestion intended to meet Professor
Perrin’s difficulty.” In summary, Lindemann’s suggestion was that the population of A*
versus A is brought into thermal equilibrium with B. The time for this equilibrium to be
established is directly related to the pressure of [B], When [B] is high enough, this
equilibrium is rapidly reestablished after A* —>P. In this case, the rate limiting step, for a
given temperature, is governed by kb. On the other hand, if [B] is small, thermal
equilibrium is established slowly compared to A* —> P, then the rate limiting step is
governed by the formation of A*, or ka[B].
In 1927-8, Rice, Rampsberg, and Kessel, proposed more detailed theories for
unimolecular reactions [52]. Their combined theory often goes by the eponym, RRK. In
summary, RRK theory states that for a unimolecular reaction to occur certain critical
vibrations and rotations must be excited. These critical modes are treated statistically by
attempting to answer the question: what is the probability that collision will result in an
A* that has the critical modes excited? Later, Marcus added his thoughts, “I blended
statistical ideas from the RRK theory of the 1920s with those of the transition state theory
of the mid-1930s [53].” And, the theory is sometimes called RRKM in its complete form.
28
For the experiments in this dissertation, the Lindemann-Hinshelwood treatment is
adequate. Williamson et al recently conducted experiments on the unimolecular thermal
electron detachment of the azulene anion [9]. These experiments were conducted at
Montana State University on the same PHPMS used for the experiments in this
dissertation. Williamson found that the transition for low pressure to the high pressure
region occurred at about I Torr. This result is unfortunate because the PHPMS has a
pressure range from about 1-5 Torn Therefore, ideal experiments, like comparing I, 10
and 100 Torr results, could not be conducted. To determine if our experiments were in
the high pressure limit, data from buffer gas pressures varying from 1-3 Torr were
frequently compared. Because the measured rate constants in Chapter 2 and the
equilibrium constants in Chapter 3 did not change at different buffer gas pressures, we
can infer the experiments were conducted in the high pressure limit.
Comparing Reaction 1-1 with Reaction 1-7, it useful to ask how do Icec and kdet
relate to ki, k _i, kz, and k _z in the high pressure limit? The detailed answer to this
question is contained in Appendix A. The final results are presented here as Equations 124 and 1-25.
( 1- 24)
( 1- 25)
29
Excitation and Thermal Electron Detachment
In a 1999 personal correspondence regarding thermal electron detachment. Miller
noted: “I’ve been thinking about this off and on for a long time with no obvious
answer...for molecules whose Icec and EA are known, isn’t the Icdet rate preordained by the
thermodynamics equations? [54].” Indeed, the laws of thermodynamics govern the ratio
of the electron capture rate constant and the thermal electron detachment rate constant,
but, the magnitude of these individual terms is independent of thermodynamic quantities
like electron affinity.
In 1985, Kebarle proposed an equation for predicting thermal electron detachment
rate constants as a function of temperature, EA, entropy upon negative ionization, and Icee
[55]. As will be shown below, Kebarle’s equation contains an assumption that may be in
error. More recently, Miller proposed an equation using different assumptions [17, 56].
Kebarle’s equation is based upon transition state theory as in Equation 1-26:
K=
(1-26)
where K is the equilibrium constant for Reaction 1-1, Q m- and Q m are the partition
functions for M' and M respectively, EA m is the electron affinity for M, R is the molar
gas constant, and Qe is the partition function for an electron. Kebarle treats the electron as
a monatomic gas with ± { spin degeneracy. Therefore Qe is given by twice the
translational partition function as in Equation 1-27:
30
(1-27)
where V is a volume term (I m3 for standard conditions), and Ine is the mass of an
electron. Kebarle’s assumption comes when he states Equation 1-28:
—e R
where A S 0ni is defined as the entropy of negative ionization given as ( S 0anIon -
(1-28)
S 0neutraI).
Kebarle asserts this to be true because of the microcanonical treatment in Equation 1-29:
S —kin
(i-29)
where £2 is the number of microstates in a given macrostate. Although Q is related to the
partition function, they are not the same thing, and the more appropriate relationship,
often referred to as the canonical ensemble; is given in Equation 1-30 [57]:
A
= - k T InQ
(i-30)
where A is the Helmholtz free energy, and Q is the partition function for the whole
system (i.e. Q = Q transhtional + Qelectronic + Qvibratiotal + Q rotationai )■Therefore, the relationship
in Equation 1-28 could be given as Equation 1-31:
=
e
RT
(1-31)
31
where AA is defined (A0anIon - A0nanw)- Kebarle then invokes the simple relationship of
forward and reverse rate constants to K as in Equation 1-32:
K
=
^ sl
(1-32)
Combining Equations 1-26, 1-27, 1-31, and 1-32 gives a canonical form of Kebarle’s
work as Equation 1-33
''A A t-E A * , X
J
K"det = ^ S x l ,Ow 15T h e
(1-33)
where Icec is given in cc sec"1 and kjet is calculated in sec"1. The microcanonical form of
Kebarle’s equation with ASin instead of AAnj in the exponential is shown here as Equation
1-34.
z AS",- . EAm ^
k,_.
‘'det = 4.SxlOl5TJkece
(1-34)
This equation has been used at least twice in the literature [58, 59], and it will be used in
this dissertation so data can easily be compared with the previous publications.
Miller’s treatment starts with the familiar relationship in Equation 1-32 and then
adds another basic concept in Equation 1-35.
AG0 = - R T l n K
= A H 0- T A S 1
(1-35)
Miller combines the relationships in Equations 1-32 and 1-35 to develop Equation 1-36:
32
273.15
kdet ~ ^ecL 0
AS (Ht -H0).
k
(1-36)
T
where L0 is Loschmidf s number (2.687 x. IO19 cc'1); EA is the electron affinity; AS
entropy change as outlined below, and (Ht -Ho) is a temperature correction for the
enthalpy term as defined below.
The AS entropy term is given below in Equation 1-37.
AS
= S anion
’neutral - S electron
(1-37)
The entropy of the electron is estimated by treating it as a monatomic gas with ± ^ spin
degeneracy and applying the Sackur-Tetrode equation as in equation 1-38 [17, 60]:
S aelecmn = k l n [
2 ( 2 e k T )* { i n
m e ^ p i p 0It 3) 7]
(1-38)
where e is the base of the natural logarithm, Hie is the mass of an electron, and p0 is a
reference pressure (101.3 kPa). The S anion - S neutral portion of the equation can be
calculated with ab initio methods or estimated with commercially available software. For
the experiments in Chapter 2, reported entropy terms were adjusted until the prediction fit
the data.
The (Ht -Ho) term in Equation 1-36 is given here as Equation 1-39.
H t -H
t
=
J c ^ f a n i o n ) d T - J c p( n e u t r a l ) d T - j c p ( e l e c t r o n ) d T
(1-39)
The anion and neutral terms are approximately the same. Therefore, Equation 1-39 can be
simplified to just the electron term which is given in Equation 1-40 [17, 61].
33
f
H t -H
5
( 1.40
\C p (e le c tr o n )d T = - — k T
t
Using the values in Table 1-2, thermal electron detachment rate constants can be
calculated as a function o f tem perature for both the M iller and Kebarle Equation as in
Figure 1-9.
Table 1-2. Therm odynam ic and kinetic properties of a typical molecule
Property
Value
Electron A ffinity
kec
Sanion — Sneutral
Equation
0.68 eV = 65.6 kJ m o l1
Both
2.23 x IO 9 cc sec'1
Both
+39 J m o l1 K 1
M iller
-3 J m o l1 K 1 (-0.71 e.u.)
A S ni
Kebarle
800 -I
600 400 -
200
-
Temperature
(C )
Figure 1-9. The solid line represents the values predicted by the Equation 1-36 (Miller);
the (+) represents values predicted by Equation I -34 (Kebarle).
34
As can be seen in Figure 1-9, both predictive equations could be used to model
the same data. Ultimately, Miller’s work in Equation 1-36, may prove more useful
because of the possible confusion regarding the microcanonical treatment of Kebarle’s
work in Equation 1-34.
Methods for Measuring Resonance Electron Capture Rates
Electron capture rates can be studied under single or multiple collision conditions
[62]. Examples of single collision experiments include electron transfer from Rydberg
atoms, rare gas photoionization, reversal electron attachment, and crossed beam studies.
Theses types of experiments involve the sample molecule interacting only with the
incoming electron or Rydberg atom. On the other hand, multiple collision experiments
involve the sample molecule reacting first with the incoming electron followed by
numerous collisions with a third body. The PHPMS is a multiple collision experiment.
Other examples of multiple collision experiments include electron swarm, Cavelleri
electron density sampling, flowing afterglow, and pulsed radiolysis.
The Flowing afterglow with Langmuir Probe (FALP) has been used to
successfully measure dissociative electron capture rates for some halogenated alkanes as
well as resonance electron capture for SFg and CU [36, 37, 62-68].
In 1994, Knighton et al reported an innovative experiment using the PHPMS to
measure electron capture rates for many substituted nitrobenzenes [20]. This study
involved using a reference compound, R, and comparing it to the sample molecule, S, as
in Equations 1-41 - 1-45:
35
kpt
R + e" - R"
(1-41)
ks
S + G -
4
«
S
**[R]
Is- « k s[ S]
(1-42)
(1-43)
(1-44)
(1-45)
where Ix is the integrated mass spectrometry signal for compound X. Knighton used CH3I
for his reference compound. There are two requirements for a good reference compound.
First it must have a well studied and accepted electron capture rate constant. Second, it
must not react with the sample molecule in any way (i.e. no transfer of electrons from
anions to neutrals). CH3I has been well studied by FALP [64]. Further, the EA for CH3I
is ~ 0.11 eV, well below the substituted nitrobenzene sample molecules which have EAs
ranging from 0.92- 2.00 eV [20]. CH3I undergoes dissociative electron capture to form I",
and atomic iodine has an electron affinity ~ 3.0 eV [69]. The electron affinity differences
mean that no ion molecule electron transfer reactions Should happen for the molecules in
this experiment.
36
The experiments in Chapter 2 of this dissertation follow the same competitive
model as Knighton et al used. However, SF6 was used as the reference compound for
reasons that are explained below.
The Suitability of SF6 as a Surrogate
SF6 is a truly unique molecule. It is thought to almost perfectly capture s-wave
electrons [23]. It is a gas at room temperature, and experimentally, it is easy to work with.
As a result, SF6 has been extensively studied. Its electron capture rate constant is well
known, ~ 2.6 x IO"7 cc sec"1 [24, 26, 37, 38, 63, 65, 70-72]. Its BA, «1.0 eY, is also well
studied [73]. Also SF6 seems to have a barrier to electron transfer reactions. This
behavior was first reported by Grimsrud in 1985 [74], In this PHPMS study, time profiles
of SF6' were compared to time profiles of anions from various molecules with higher EAs
than SF6. Although electron transfer from SF6" to the higher BA molecule should occur
on every collision (i.e. kc = 2.0 x IO"9 cc sec'1), Grimsrud noted these rate constants were
< 10"13 cc sec"1. Recently, Neumaier reported, in a personal communication, the reverse
of this phenomenon [75]. Anions and dianions from fullerenes like C76, C7g, and C60 were
trapped in an ion cyclotron mass spectrometer (ICR) with approximately 2 x 10"8 mTorr
SF6. The electron transfer to SF6 was thermodynamically favored in each case, but no
SF6" was seen. Reaction 1-46 is a typical example.
C782"+ SF6-
C781"+ SF6" AH = -0.8 eV
(1-46)
37
If the fullerene ions were accelerated in the ICR, the electron transfer reaction would
eventually take place. But, the threshold was at about 2.0 eV.
Conclusion
Understanding the experiments that follow requires knowledge of several key
points. First, the PHPMS was explained with special emphasis on how data are collected
and examined. The plasma formed in the ion source complicates the behavior of ions and
electrons as they diffuse to the walls. A discussion of ambipolar diffusion and the
different types of plasmas, (i.e. electron dominated and ion dominated) was presented to
clarify these issues. Various researchers have examined electron capture and the
formation of long lived intermediates. Their work has been summarized to provide an
introduction to the vocabulary and formalisms of this field. Finally, the role of the buffer
gas and the high pressure limit has been explained both from a historic and a
mathematical perspective. The next two chapters will present original research. In
Chapter 2, electron capture rate constants and thermal detachment rate constants have
been measured for several low EA molecules. Chapter 3 will answer a 30 year old
mystery about why anions from these molecules are not seen at atmospheric conditions.
38
RESONANCE ELECTRON CAPTURE AND THERMAL DETACHMENT
Resonance electron capture and thermal electron detachment rate constants have
been successfully measured of benzophenone, quinoxaline and anthracene. Typical of
chemistry literature, these measurements will be presented in two sections: a)
experimental, and b) results and discussion. The benzophenone discussion will chronicle
the faulty assumptions that were overcome to successfully understand the data. The main
results with less detailed discussions will follow for quinoxaline and anthracene. Finally,
experiments on quinazoline will be presented showing the types of problems caused by
background O2. The quinazoline/02 problem will lead nicely into the experiments in
Chapter 3.
Benzophenone
Benzophenone has a reported electron affinity of 0.61 eV [76]. Throughout this chapter
the benzophenone anion will be abbreviated as BP". Benzophenone has a mass of 182
amu, and its structure is depicted in Figure 2-1.
O
Figure 2-1. The molecular structure of benzophenone.
39
Experimental
The PHPMS was used for all experiments. The gas handling plant was prepared
by injecting a solution of benzophenone in toluene. Injection amounts ranged from 1.31 x
IO"5 - 1.73 x IO"4 moles of benzophenone which resulted in mixing ratios of 64.1 - 843
ppm in the buffer gas. The Clasius-Clapyeron equation was used to estimate vapor
pressure limits for benzophenone in the gas handling plant, transfer line, and the ion
source; injection amounts were limited to stay below the vapor pressure limit. Pure SFe
was injected in amounts ranging from 2.10 x IO"7 - 1.11 x IO"4 moles which resulted in
mixing ratios of 1.7 - 565 ppm in the buffer gas. SFe injection amounts were generally
governed by the goals of each individual experiment.
Data were collected at source pressures ranging from 2-4 torn All experiments
involved taking mass spectra to determine which ions were present. Then, time profiles
were taken for each ion present. Two types of experiments were conducted. The first type
involved setting the ion source temperature and the benzophenone concentration and
adding successive SFg injections. The second type involved setting the SFg and
benzophenone concentrations and varying the ion source temperature.
Results and Discussion
The initial goal of this experiment was to duplicate the Knighton study [20]
(detailed in Chapter I), and mass spectra were recorded for a mixture of benzophenone
and SFg as in Figure 2-2.
40
2000 i
1500«
1000 «
2500i
2000
«
1500*
1000
«
3000i
2500*
2000*
1500«
1000 «
M ass (m /z)
Figure 2-2. Mass spectra of benzophenone anion and SF6" at 3 torn (A) T = 35°C, (B) T =
50°C, and (C) T = 75°C. The relative number density of the neutral benzophenone and
SF6 is the same in each spectra, but the temperature is different. Concentration of
benzophenone = 147 ppm , SF6 = 8.3 ppm (as a fraction of the buffer gas mixture).
41
Apparent Loss of BP to SF^. As mentioned in Chapter I, electron capture rates
are believed to be insensitive to temperature in the range of this experiment. Therefore,
TYl
the relative decrease in the benzophenone peak ( — = 182) compared to the SF6 peak
m
(— = 146) as temperature increases is indicative of electron transfer as in Reaction 2-1.
^transfer
BP +SF6- B P + SF6-
(2-i)
Time profiles often reveal if an ion is being consumed or produced, and Figure 2-3
contains the time profiles for the ions in Figure 2-2 (B).
Time (msec)
Figure 2-3. Normalized time profile of BP (+) and SF6 (o) taken from data collected
under the same conditions as the mass spectrum in Figure 2-2 (B). These data strongly
suggest that BP " is being lost while SF6" is being produced.
42
Recall from Chapter I that SF6 was specifically chosen as a surrogate because of
independent experimental evidence that it does not undergo Reaction 2-1. However, the
time profiles in Figure 2-3 clearly indicate BP is being lost while SF6: is being produced.
In order to verify if Reaction 2-1 was actually occurring or not, standard kinetic
experiments were conducted by adding successive SF6 injections to a given
benzophenone concentration at a set ion source temperature. The first order loss of BP
was calculated for each SF6 concentration, and the second order RtranSfer rate constant was
calculated as in Equation 2-2:
k
^transfer
(2-2)
where k is the psuedo first order loss of BP". For a brief mathematical review of psuedo
first order kinetics, see Appendix B. The results, in Figure 2-4, indicate that the loss of
BP to SF6 is not a second order process. If it were second order, RtranSfer would be
constant for all SF6 concentrations. So, what is happening to the BP ? As shown in
Reactions 1-1 and 1-7, electron capture is a reversible process. A detached electron can
be recaptured by benzophenone (and re-detached ad infinitum), or it can be captured by
SF6. SF6' does not appear to detach electrons in the temperature range of these
experiments. Therefore, once a detached electron is scavenged by SF6, it is lost.
43
1.2E-09
9. OE-10
6. OE-10
3. OE-10
0.0E+00
0.0E+00 3.0E+12 6.0E+12 9.0E+12 1.2E+13
SFg Number Density
Figure 2-4. The calculated 2n<l order rate constant for the direct electron transfer reaction
BP + SF& ^ BP + SF6 as a function of SF6 number density. If this process were actually
occurring, the 2nd order rate constant would not decrease with increasing SF6 number
density.
In this case, Reaction 2-3 correctly describes the chemistry:
Icdel
ka[SF6]
BP ^ BP + e
SF6
(2-3)
kec
where ka is the electron capture rate constant for SF6.
Time Profile Data and Differential Equations Model. The thermal electron
detachment problem just described makes the mass spectra data almost worthless for
44
calculating the electron capture rate constant for benzophenone. However, examining the
time profiles revealed a very interesting result as in Figure 2-5.
1.00 -I
0.80 -
0.20
- 0.80
£ 0.60
i
0.40
0.20
-
1.00 -I
0.80
0.60 0.40 0.20
Time (msec)
Figure 2-5. Time profiles of BP" (+) and SF6" (o), taken under the same conditions as the
mass spectra in figure 2-2. (A) T = 35°C, (B) T = 50°C, and (C) T = 75°C. Although the
BP" decays away faster at higher temperatures, the relative initial populations of BP and
SF6"is not affected by temperature.
45
The initial concentrations of BP" and SF6' are essentially not affected by the ion source
temperature. But, the rate of Reaction 2-3 is affected by the temperature. At 35°C, the
BP is not depleted after 40 msec. At 50°C, the BP population is depleted after 30 msec,
and at 75°C, the BP population is depleted after 10 msec. This result is encouraging
because the initial concentrations might be useful for calculating the electron capture rate
constant using Equation 1-45.
To determine whether the result in Figure 2-5 was due to actual chemistry or just
an artifact of the PHPMS instrument, the experiment was modeled with the IBM Kinetic
Simulator. This software package is a stochastic differential equation solver. The
diffusion rates for SF6" and BP" were assumed to be the same. Table 2-1 shows the
electron capture and thermal detachment rate constants required for the IBM Kinetic
Simulator to match time profiles from the experimental data taken at 2 and 3 Torr and the
temperatures indicated. The benzophenone concentration was ~ 147 ppm, and the SF6
concentration was ~ 8.3 ppm (as a fraction of the buffer gas mixture).
Table 2-1. Parameters used for the IBM Kunetic Simulator
Temperature (0C)
kdet (sec"1)
kec (cc sec"1)
40
50
. 1.50 x 10"8
46
120
1.50 x 10"8
51
200
1.50 x 10"8
55
285.
1.50 x 10"8
46
Table 2-1 continued
Temperature (0C)
k det
(sec"1)
kec
(cc sec"1)
60
400
1.50 x 10"8
65
575
1.50 x 10"8
69
750
1.50 x 10"8
75
1100
1.50 x 10"8
In summary, the IBM kinetic simulation program showed that the initial
populations of each ion were directly proportional to the product of the assigned electron
capture rate constant and the number density of the neutral species. Therefore, the
information from the early part of the time profiles about relative concentration can be
used in Equation 1-45 to calculate the electron capture rate constant for benzophenone.
The other lesson learned from these simulations is that the disappearance of BP is related
to the thermal electron detachment rate constant and the relative amounts of SFe and
benzophenone that are present.
As outlined in Chapter I, in the first few msec of the ion time profiles, the plasma
in the ion source is electron dominated. During this time, the negative ions are essentially
bottled in the ion source (i.e. kdiff = 0) and the electrons are diffusing to the wall.
Unfortunately, this early region in the time profile contains the data needed to calculate
the electron capture rate constant. To overcome this problem, the initial populations of
BP and SFe' were extrapolated from logarithmic time profiles as in Figure 2-6.
47
12.0 -I
Time (msec)
Figure 2-6. Logarithmic time profiles of BP (+) and SF& (o) demonstrating the technique
for extrapolating the initial number densities of each ion.
Experiments were then conducted at various temperatures, pressures and concentrations
of SFe and benzophenone. The measured electron capture rate constants are in Figure 2-7.
3.0E-08 2.0E-08 -
J
1.0E-08 ■
0.0E+00
Pressure (Torr)
Temperature (C)
Figure 2-7. Measured electron capture rate constants for benzophenone. These data were
collected at various temperatures, pressures, and relative concentrations of SF^ and
benzophenone.
48
The average resonance electron capture rate for benzophenone is 1.65 x IO 8 cc
sec'1. Note that the electron capture rate constant appears to be independent of
temperature. It is also independent of pressure which indicates the experiments are being
done in the high pressure limit. Further, Icec values calculated from normalized time
profile data were essentially the same as the values shown in Figure 2-7.
Failure of Model at Low Number Densities. The next experimental goal was to
measure the thermal electron detachment rate at various pressures. Appendix C contains a
detailed mathematical workup of the differential equations that describe BP and the
electrons. Equation 2-4 is the result of integrating and algebraically manipulating the
differential equation. Also, the following three assumptions were invoked: I) the
electrons are in a steady state, 2) no electrons are being lost to diffusions, and 3) no
impurities are present.
Mn[ B P ]
kJeA
At
[ M ]
Kc [SF6]
|
I
kdet
(2 -4 )
The term in the left hand side of this equation has experimental meaning. On a
A ln [ B P -]
logaiithimic plot of the time profile,
^
represents the slope of the line as in
Figure 1-3. This observed first order loss is the result of BP' both diffusing to the wall
and being lost to thermal electron detachment followed by electron scavenging by SFg as
49
] ,
A ln [ B P
shown in reaction 2-3 above. So, the term
.
' ^diff , represents the observed
ISt
chemical loss of BP'. To make Equation 2-4 less cumbersome the term ki0ss has been
invented as in Equations 2-5 and 2-6.
Z
ln[ BP ]
At
V
A
\
-L
,A
Kc
- k loss
J
(2-5)
]
[ SF6]
kiet
(2-6)
This equation is convenient because it should be linear for a given temperature. Once the
y intercept is established, it can be used in conjunction with the slope to calculate the
electron capture rate constant, Icec. Remember, ka (2.6 x IO'7 cc sec"1) is the electron
capture rate constant for SFg. Typical results are shown in Figure 2-8. The prediction line
was created by using the average Icee value measured in Figure 2-7, and estimations of kdet
from the IBM Kinetic Simulator program. The data in Figure 2-8 are interesting for two
reasons. First, at higher relative SFg concentrations,
[BP]
I SF6J
^ 20, the data match the
prediction line. However, at lower SFg concentrations the observed kioss is higher than
50
1.0E-02
8.0E-03
6.0E-03
4.0E-03
2.0E-03
0
10
20
30
40
50
60
70
[BP]/[SF6]
Figure 2-8. A plot of Equation 2-4. Data were collected at 60°C and 2 -3 Torr at various
relative SF^ and benzophenone concentrations as indicated on the graph. The line is
predicted for Itec = 1.65 x Kf8cc sec 1and kdet = 450. Benzophenone number densities (as
a fraction of the buffer gas mixture) (o) 212 ppm, (A) 423 ppm, and (x) 843 ppm. SFe
concentrations can be deduced by dividing the X axis value into the given benzophenone
concentration.
predicted by Equation 2-6 (note that in Figure 2-8, the y axis is the inverse of ki0ss).
Second, the data represented by (A) and (x) represent experimental conditions where the
total number densities of benzophenone and SF& are 2 times and 4 times greater than the
experiment represented by (o). The (A) and (x) higher number density data are closer to
the prediction line than the (o) lower number density data. This result suggests that one or
more of the assumptions listed above are incorrect. For example, what if some of the
electrons were being lost to diffusion? In this case, ki0ss would be higher than predicted
51
especially in experimental conditions where the number density of electron capturing
species was low.
The New Model with Electron Diffusion. While the linear model does not appear
useful, for presenting data, it was very useful for revealing this faulty assumption.
Equation 2-7 is the complete integrated equation describing BP" in time.
Z
ln [ M
]t —
Ket - K ijf ~ K n J x ]
7n
t + ln [ M
(2-7)
This equation is also derived in Appendix C and involves a steady state treatment of the
electrons. Equation 2-7 can be used to predict the slope of the logarithmic plot for a BP
time profile. The only unknown parameters are the electron diffusion rate, kediff, and the
concentration of high EA impurities, [X].
Attempts to rigorously measure the electron diffusion rate constant where not
successful. However, using data from experiments with benzophenone and SFg at 60°C,
and Equation 2-7, kediff was calculated for each data point. Figure 2-9 displays these
results in a historgram. At 4 Torr and 60°C, kediff is probably between I x IO4 and I x IO5
sec"1.
Figure 2-10 is a comparison of predicted values for the BP versus actual
measured values. These experiments were done over a wide range of SFg and
benzophenone concentrations.
52
40 T
35 ••
30 - 25 - 20
- •
15 - *
10 -•
-7.00E+05
-2.00E+05
3.00E+05
8.00E+05
Electron Diffusion Rate Constant
Reduced for 30C and 4 TonFigure 2-9. A histogram of reduced electron diffusion rate constants from experiments
with benzophenone and SFa at 60°C.
Data were collected at buffer gas pressures of primarily 2 and 3 Torr. The temperature
ranged from 35-75°C; kdet for each temperature was estimated from experiments where
the number densities of SF6 and benzophenone were high, and the linear model was
approximately correct. The data points above 750 have significantly more spread than the
points between 200 and 600. The >750 data were collected at 75°C where kdei is high. As
a result, the number of linear data points on the logarithmic time profile is limited, and
the uncertainty of measured slopes is higher than the lower temperature experiments.
The kediff value was approximated to be 5 x IO4 sec'1(at 30°C and 4 Torr). Also, an
TTl
impurity peak, probably from chlorobenzophenone, was noted at — = 217. To account
z
for this peak, [X] was estimated to be = 0.025% of the benzophenone concentration.
53
1250 i
1000 750
-
500 -
250
-
M easured Slope
Figure 2-10. A comparison of benzophenone data versus calculated values of Equation 27 when the following parameters are used: kec = 1.65 x IO 8 cc sec ', kdiff= 92.7 sec"1 (4
torr, 30C), kediff = 5 x IO4 sec"1(4 torr, 30°C). Data were collected from 35 -75°C and 2-3
Torr, kdet = 250, 450, 1200 was estimated for each data point based on previous
experiments where SFe was in great excess. Electron transferring impurity = 0.025% of
the benzophenone number density.
Predictions of Thermal Detachment. Equation 1-34 put forth by Kebarle [55], and
Equation 1-36 put forth by Miller [17, 56], can both be used to predict thermal electron
detachment rate constants as a function of temperature. Table 2-2 shows the
thermodynamic values used for the Miller and Kebarle equations to predict the data.
Figure 2-11 compares measured thermal detachment rate constants to predicted values
from both the Kebarle and the Miller equations.
54
Table 2-2. Thermodynamic and kinetic properties of benzophenone
Property
Value
Electron Affinity
kec
Sanion — Sneutral
A Sni
1200
Equation
0.61 eV = 58.9 kJ mol'1
Both
1.65 x IO'8 cc sec"1
Both
+38.5 I mol'1K 1
Miller
-3.3 I m ol1 K 1 (-.79 e.u.)
Kebarle
-
800 -
400 -
Temperature (C)
Figure 2-11. Thermal electron detachment predictions for BP using the Kebarle and
Miller equations with the parameters in Table 2-2. (o) are measured data for BP'. The
solid lines which lie almost on top of each other are the predictions from each equation.
Efforts to Stabilize BP with SiF^
Williamson et al has reported a clustering reaction between BP and Sip4 as
indicated in Reaction 2-8 [77].
BP + SiF4 - (BPeSiF4)"
(2-8)
55
Williamson suggests that this reaction lies far enough to the right that it might stop
thermal electron detachment. If this hypothesis is true, electron capture rate constants for
benzophenone could be measured by comparing the (BPoSiF^" peak ( — = 286) with the
z
SFe peak in typical mass spectra as in Figure 2-12. Comparing the spectra in Figure 2-12
A and B, shows that Sip4 does protect BP from thermal electron detachment. But,
comparing the spectra in Figure 2-12 B and C reveals that the (BPoSiF4)" peak is
relatively larger at lower temperatures. This increased signal implies that Reaction 2-8 is
not shifted completely to the right as suggested by Williamson. And while SiF4 is useful
for qualitatively detecting BP" at high temperatures, it is not useful for quantitative
electron capture rate constant determinations.
Quinoxaline
Quinoxaline has a reported electron affinity of 0.68 eV [76]. Throughout this
chapter the quinoxaline anion will be abbreviated Qx'. Quinoxaline has a mass of 130
amu, and its structure is depicted in Figure 2-13.
Experimental
The PHPMS was used for all experiments. The gas handling plant was prepared
by injecting a solution of quinoxaline in toluene. Injection amounts ranged from 1.27 x
IO'5 - 7.52 x IO"3 moles of quinoxaline which resulted in mixing ratios of 61 ppm - 36.8
ppt in the buffer gas.
Intensit
56
Figure 2-12. Mass spectra for the (BP*SiF4)" experiment. Spectrum A is at 90°C with 416
ppm benzophenone and 26 ppm SF6. Spectrum B is the addition of 18.4 ppm SiF4.
Spectrum C is the same mixture at 61°C
57
Figure 2-13. The molecular structure of quinoxaline.
The Clasius-Clapyeron equation was used to estimate vapor pressure limits for
quinoxaline in the gas handling plant, transfer line, and the ion source; injection amounts
were limited to stay below the vapor pressure limit. Pure SFe was injected in amounts
ranging from 6.98 x IO"7- 1.85 x IO"5 moles which resulted in mixing ratios of 0.3 - 94
ppm in the buffer gas. SFe injection amounts were generally governed by the goals of
each individual experiment.
Data were collected at source pressures ranging from 2-4 torn All experiments
involved taking mass spectra to determine which ions were present. Then, time profiles
were taken for each ion present. Two types of experiments were conducted. The first type
involved setting the ion source temperature and the quinoxaline concentration and adding
successive SFe injections. The second type involved setting the SF6 and quinoxaline
concentrations and varying the ion source temperature.
Results and Discussion
Similar to benzophenone, electron capture rate constants were calculated for
quinoxaline by extrapolating the initial relative concentrations of Qx' and SF6" from
58
logarithmic time profiles. Figure 2-14 shows the calculated rate constants versus pressure
and temperature.
4.0E-09
3.0E-09
2.0E-09
oo o@
1.0E-09
0.0E+00
4.5 0
Pressure (Torr)
150
200
Temperature (C)
Figure 2-14. Measured electron capture rate constants for quinoxaline. These data were
collected at various temperatures, pressures, and relative concentrations of SFo and
quinoxaline.
Similar to benzophenone, the quinoxaline experiments appear to be in the high pressure
limit, and the electron capture rate does not seem to vary with temperature. The average
electron capture rate is 2.31 x 10'9 cc sec"1. Even though the EA of quinoxaline is
approximately 0.07 eV higher than benzophenone, quinoxaline captures electrons about
seven times slower.
Equation 2-7 can also be used to compare the measured sloped of the Qx
logarithmic time profiles to predicted values as shown in Figure 2-15. As might be
expected, the same reduced electron diffusion rate constant used for the benzophenone
experiments also works reasonable well to explain the quinoxaline data.
59
800 i
600 400 -
200
-
Measured Slope
Figure 2-15. A comparison of the quinoxaline data versus calculated values of Equation
2-7 when the following parameters are used: Icec = 2.31 x 10 9 cc sec"1, kdiff= 92.7 sec"1 (4
torr, 30C), IcetJjff = 5 x IO4 sec'1 (4 torr, 30C) Data were collected from 750-145°C and 2-4
Torr, kdet was estimated for each data point based on previous experiments were SF^ was
in great excess.
The thermal electron detachment rate constants can also be calculated using the
Kebarle or Miller equations, Equations 1-34 and 1-36 respectively, with the parameters in
Table 2-3.
Table 2-3. Thermodynamic and kinetic properties of quinoxaline
Property
Electron Affinity
kec
Sanion ~ Sneutral
A S nj
Value
Equation
0.68 eV = 65.6 kJ mol"1
Both
2.31 x IO 9 cc sec"1
Both
+39 J mol"1 K 1
Miller
-3.0 J mol"1 K 1(-0.72 e.u.)
Kebarle
60
Figure 2-16 compares measured thermal electron detachment rate constants with
predictions from the Kebarle and Miller Equations.
800
i
600
■
Temperature (C)
Figure 2-16. Thermal electron detachment predictions for Qx using the Kebarle and
Miller equations with the parameters in Table 2-3. (o) are measrured data for Qx'. The
dashed line is from the Kebarle equation, and solid line is from the Miller equation
While quinoxaline captures electrons slower than benzophenone, comparing Figures 2-16
and 2-11 reveals that quinoxaline also detaches electrons much more slowly than
benzophenone.
Anthracene
Anthracene has a reported electron affinity of 0.54 eV [76]. Throughout this
chapter the anthracene anion will be abbreviated as A". Anthracene has a mass of 178
amu, and its structure is depicted in Figure 2-17.
61
Figure 2-17. The molecular structure of anthracene.
Experimental
The PHPMS was used for all experiments. The gas handling plant was prepared
by injecting a saturated solution of anthracene in toluene. Injection amounts ranged from
2.24 x 10 6 - 3.59 x IO 5 moles of anthracene which resulted in mixing ratios of 10.6 196.2 ppm in the buffer gas. The Clasius-Clapyeron equation was used to estimate vapor
pressure limits for anthracene in the gas handling plant, transfer line, and the ion source;
injection amounts were limited to stay below the vapor pressure limit. Pure SFa was
injected in amounts ranging from 5.39 x IO'8 - 7.02 x IO 7 moles which resulted in
mixing ratios of 0.3 - 3.6 ppm in the buffer gas. SFa injection amounts were generally
governed by the goals of each individual experiment.
Data were collected at source pressures ranging from 2-4 torr. All experiments
involved taking mass spectra to determine which ions were present. Then, time profiles
were taken for each ion present. Experiments involved setting the SFa and anthracene
concentrations and varying the ion source temperature.
62
Results and Discussion
Electron capture rate constants were measured in the same fashion as
benzophenone and quinoxaline. However, the logarithmic time profiles of A appeared
abnormal at higher pressure as shown in Figure 2-18.
Time (sec)
Figure 2-18. Time profiles of A at 2 Torr (o), 3 Torr ( ), and 4 Torr (A) demonstrating
the non-linear behavior as pressure increases. The full and dashed lines show indicate
how the 4 torr data might consist of time profiles from two different anions.
The time profiles in Figure 2-18, suggest that an impurity with the same unit mass as
anthracene could be present in the sample.
63
Close examination of the 4 Torr data reveals that from about 5-10 msec the logarithmic
time profile is linear with a slope given by the dashed line. After 15 msec, the logarithmic
time profile is linear with a less steep slope given by the solid line. This flattening of the
time profile lowers the extrapolated value for the initial concentration of A". Because this
flattening is more prevalent at higher pressures, the measured electron capture rate
constants appear to have a negative pressure dependence as shown in Figure 2-19.
3 .0 E -0 9 i
2 .0 E -0 9 -
1 .0 E -0 9 -
P r e ssu e (Torr)
Figure 2-19. Calculated electron capture rate constants calculated from the logarithmic
time profiles for anthracene.
Efforts were made to identify the phantom impurity. High resolution mass spectra
revealed a peak at 179 amu, but this peak is explained by the carbon 13 isotope. The
resolution of the PHPMS quadrupole mass spectrometer was not good enough to conduct
more meaningful experiments.
As will be shown in Chapter 3, it is also possible for A to react with O2 and H2O
to form various cluster products. These clustering reactions are reversible. So, another
explanation for the non-linear behavior in Figure 2-18 is that the reverse cluster reactions
are replenishing the A" population. However, mass spectra for these experiments showed
very low intensity signals for these cluster products.
In summary, extrapolating the initial concentrations of SF6" and A from the
logarithmic time profiles does not appear to work for anthracene like it did for
benzophenone and quinoxaline. To solve this problem, normalized time profiles for
benzophenone and quinoxaline were compared with the logarithmic time profiles. And
for both of these molecules the electron capture rate constants calculated from the
logarithmic time profiles were in good agreement with those calculated from normalized
time profiles. Normalized time profiles for anthracene from the same data that lead to
Figure 2-19 were used to calculate the electron capture rate constants shown in
Figure 2-20.
4.0E-09 ■
O
3.0E-09 2.0E-09 -
0
O
U
1.0E-09 -
O
:
9
&
&
0.0E+00 1.0
2.0
3.0
4.0
Pressure (Torr)
20
40
60
80
Temperature (C)
Figure 2-20. Measured electron capture rate constants for anthracene. These data were
collected at various temperatures, pressures, and relative concentrations of SF6 and
anthracene.
65
As can be seen by comparing Figures 2-19 and 2-20 the normalized time profiles
lead to more behaved electron capture rate constants. The average Icec = 2.09 x IO'9 cc
sec"1. As with benzophenone and quinoxaline. Equation 2-7 can be used to predict the
slope of the logarithmic time profiles for a given set of parameters as in Figure 2-21.
a iooo
250
500
750
1000 1250
1500
Measured Slope
Figure 2-21. A comparison of anthracene data versus calculated values of Equation 2-7
when the following parameters are used: kec = 2.09 x IO 9 cc sec"1, kdiff= 92.7 sec"1(4 torr,
30C), kediff = 5 x IO4 sec 1 (4 torr, 30C) Data were collected from 40-81°C and 2-4 Torr,
kdet was estimated for each data point similar to previous benzophenone and quinoxaline
experiments.
Once again the same electron diffusion term used for the previous molecules
explains the anthracene data.
66
The thermal electron detachment rate constants can also be predicted using the
Kebarle or Miller equations. Equations 1-34 and 1-36 respectively, with the parameters in
Table 2-4.
Table 2-4. Thermodynamic and kinetic properties of anthracene.
Property
Electron Affinity
kec
Sanion —Sneutral
ASnj
Value
Equation
0.54 eV = 52.1 kJ mol"1
Both
2.09 x IO"9 cc sec'1
Both
+42 J mol"1 K 1
Miller
0.0 J mol"1 K 1 (0.0 e.u.)
Kebarle
Figure 2-22 compares measured thermal electron detachment rate constants with
predicted values from the Kebarle and Miller equations using the parameters in
Table 2-4. One caution should be noted regarding the measured thermal electron
detachment rate constants. These data are collected from the slope of the logarithmic time
profiles. As mentioned earlier, these data, especially at high pressures, are suspect for
anthracene. To be certain, the measured detachment rates are at least as high as they
appear in Figure 2-22; they could, in fact, be higher.
Ouinazoline
Quinazoline has a mass of 130 amu, and its EA has been reported at 0.56 eV [76].
Throughout this section the quinazoline anion will be abbreviated Q z'.
67
1500 i
1000 -
500 -
Temperature (C)
Figure 2-22. Thermal electron detachment predictions for A using the Kebarle and Miller
equations with the parameters in Table 2-4. (o) are measrured data for A". The dashed
line is from the Kebarle equation, and solid line is from the Miller equation.
It is the meta-substituted isomer of quinoxaline as shown in Figure 2-23.
Figure 2-23. The molecular structure of quinazoline.
Experimental
The PHPMS was used for all experiments. The gas handling plant was prepared
by injecting a solution of quinazoline in toluene. Injection amounts ranged from 1.38 x
10'5 - 2.27 x IO 5 moles of quinazoline which resulted in mixing ratios of 66 -102 ppm in
68
the buffer gas. The Clasius-Clapyeron equation was used to estimate vapor pressure
limits for quinazoline in the gas handling plant, transfer line, and the ion source; injection
amounts were limited to stay below the vapor pressure limit. Pure SFe was injected in
amounts ranging from 1.05 x IO"7 - 9.06 x IO"7 moles which resulted in mixing ratios of
0.1 - 0.9 ppm in the buffer gas. SFe injection amounts were generally governed by the
goals of each individual experiment.
Data were collected at source pressures ranging from 2-4 torr. All experiments
involved taking mass spectra to determine which ions were present. Then, time profiles
were taken for each ion present. Experiments involved setting the SFg and quinazoline
concentrations and varying the ion source temperature.
Results and Discussion
Quinazoline immediately presented a problem as indicated in Figure 2-24. In this case,
there are two peaks of almost equal intensity, the Qz' peak at 130 amu and another peak
at 162 amu. From the experiments in Chapter 3 it will be shown that 162 amu ion is the
(QzoOa)" cluster. However, no oxygen was added to produce this spectrum. The PHPMS
is fitted with Oa and HaO traps, but background levels of these impurities are probably 15 ppm. This small amount of Oa is responsible for signal at 162 amu as shown in
Reactions 2-9.
Q z
+ O 2 ^
( Q Z 0 O 2) ^
?
(2- 9)
69
'53 600
0
50
100
150
200
250
M a s s (m /z )
Figure 2-24. Typical mass spectrum for quinazoline taken at 3.0 Torr and 35°C.
Quinazoline concentration was 102 ppm (as a fraction of the buffer gas).
The question mark in Reaction 2-9 is important. Again Chapter 3 will address how
(Qz-O2)' reacts with H2O. Further, the O2 may not thermally detach electrons, or it might
detach them much slower than Qz " alone. The normalized time profiles showed that
Reaction 2-9 did not reach a steady state equilibrium condition in the time scale of these
experiments.
While Reaction 2-9 is obviously a complication, it occurs after the Qz " is
formed. So, it is conceivable that the initial concentrations of SF^ and Qz " could still be
used to measure the electron capture rate constant. Figure 2-25 shows measured electron
capture rate contants from both the logarithmic and the normalized time profiles. The
70
1 .2 E - 0 8 i
%
8 .0 E - 0 9 -
4 .0 E - 0 9 ■
0 .0 E + 0 0
Pressure (Torr)
Figure 2-25. Measured electron capture rate constants for quinazoline. (o) were
calculated from logarithmic time profiles, (x) were calculated from normalized time
profiles.
results are interesting for two reasons. First the values calculated from the normalized
time profiles (average ~ 7.67 x IO'9 cc sec'1) are about 3.4 times larger than the values
calculated by the logarithmic time profiles (average = 2.23 x IO9 cc sec'1). Second, the
values calculated with logarithmic time profiles show a linear decrease with temperature
(R2= 0.94). Treating the quinazoline data the same as the anthracene leads us to infer that
the electron capture rate constant is = 7.67 x IO'9 cc sec'1. However, uncertainty about the
effects of the O2 clustering reaction leads to greater uncertainty in this measurement than
with the previous molecules.
Thermal electron detachment is also complicated by the O2 clustering reaction.
Although not explicitly stated in the literature, it is generally accepted that a cluster
anion, like (QzeO2)', does not directly detach electrons. Rather, it falls apart and then the
molecular anion, Qz', detaches the electron. If this premise is true, the clustering reaction
would have the effect of acting as a reservoir of protection against thermal detachment.
and, measured thermal electron detachment rate constants would be only a fraction of
their true value. If the first half of Reaction 2-9 reached equilibrium, the steady state
concentrations of Qz “ and (QzeO2) could be used with the equations in Appendix C to
adjust the measured values. Unfortunately, Reaction 2-9 does not reach equilibrium in the
time scale of these experiments.
Measured thermal electron rate constants shown here in Figure 2-26 represent a
minimum for each temperature.
1200
-
O
1000 -
O
-T s 800 a>
OO
o:
O
600 -
"ti
^
400 -
200
O
O
O
o
I
-
0 -
35
45
55
65
75
Temperature (C)
Figure 2-26. Observed loss of Qz ~ to thermal electron detachment. The measured values
have not been corrected for the effects of the O2 clustering reaction.
The actual rate constants could be 2-5 times bigger.
72
Conclusion
Electron capture and thermal electron detachment rate constants were measured
for benzophenone, quinoxaline and anthracene. In addition, the loss of electrons by
diffusion was determined to be important in making these measurements, and the electron
diffusion rate, in the plasma conditions present in the PHPMS, was determined to
approximately 300 times greater than the diffusion rate for negative ions. Once the
electron capture rate constant is known, predictive equations for thermal electron
detachment, previously purposed by Kebarle and Miller can be used to calculate the
thermal electron rate constant as a function of temperature. Table 2-5 summarizes the
findings for each compound.
Table 2-5. Summary of kinetic and thermodynamic properties for each molecule
Compound
Benzophenone
kec
(cc sec'1)
■EA (eV)
Sanion -Srmetra]
A S ni
(J K 1mol"1)
(J K 1mol"1)
1.65 x IO 8
0.61
+38.5
-3.3
Quinoxaline
2.31 x IG'9
0 .6 8
+39
-3.0
Anthracene
2.09 x IO"9
0.54
+42
0.0
Quinazoline
2 .0 - 9,0 x IO"9
0.56
not reported
not reported
Both anthracene and quinazoline produced complicated time profiles. Reactions
with O2 were demonstrable problematic for quinazoline, and 0% may have contributed to
the abnormal time profiles for anthracene. As will be explored in detail in Chapter 3, O2
and another ever present impurity, H2O, having been causing problems in the detection of
anions from low EA molecules for the past 30 years.
74
FAILURE TO DETECT THESE ANIONS AT ATMOSPHERIC CONDITIONS
Introduction
In the environmental, biomedical and forensic sciences, an ever-increasing need
exists for the trace detection and analysis of specific substances in complex samples.
Everyday applications include finding a particular pollutant in a sample of lake water,
determining Freon concentrations in the atmosphere, and searching for toxins in body
fluids.
Some of the most promising methods that have been developed have been based
on the gas phase negative ionization of compounds by the attachment of thermal-energy
electrons as symbolized by Reaction 3-1[2, 78-81] :
e~ + M ^ M "
(3-i)
For compounds with a large positive EA, > 1.0 eV, the second order forward rate
constants, Icec, is often very large, on the order of IO"7 cc sec'1 [20]. This fast reaction
accounts for the extraordinarily high sensitivity that can be obtained by methods based on
REC. The most common instruments are the electron capture detector (ECD) [82-84], the
ion mobility spectrometer (IMS) [85], the atmospheric pressure ionization mass
spectrometer (APIMS) [86], the ion mobility spectrometer with detection by mass
spectrometry (IMS/MS) [87] , and the negative chemical ionization mass spectrometer
(NCIMS) [79, 80].
75
Two problems have been encountered with these methods that rely on electron
capture and the formation of negative ions. First, as indicated by the double arrow.
Reaction I is reversible. This elementary step, also well detailed in the previous chapters,
is called thermal electron detachment (TED). It is a particularly undesirable reaction
since it completely destroys the molecular anions of interest and the associated response
to the analyte, M. The rates of TED reactions increase strongly with increased
temperature and decreased EA of M (this relationship is mathematically described by
equation 1-32 and 1-34 in Chapter I) so that, at commonly used ion source temperatures
of about 150°C or greater, TED typically becomes unacceptably fast for compounds
having EA values of less than about 0.78 eY [55, 58, 79]. Second, some classes of
compounds react with trace levels of oxygen and water (both are commonly present in
the buffer gas of an ion source). For example, the molecular anions of halogenated
aromatic hydrocarbons are consumed in reactions with oxygen [88]. Molecular anions of
perfluorinated aliphatic hydrocarbons also undergo fast side reactions with trace levels of
water [89].
A recent report [77] demonstrated that the detrimental effects of the TED reaction
on the REC mass spectra of some low electron affinity compounds (such as
benzophenone) obtained at relatively high ion source temperatures could be overcome by
the intentional addition of small amounts of silicon tetrafluoride to the ion source buffer
gas. Due to a strong Lewis acid-base interaction between SiF4 and the molecular anions
of Iow-EA compounds that have a sterically unhindered Lewis base site, a strong
molecular anion - SiF4 complex is formed which prevented the TED reaction for low EA
76
compounds. As explained in the benzophenone section in Chapter 2, the Sip4 reaction
appears to be reversible. Therefore, TED is not stopped completely. Rather, a fraction of
the parent anion is protected from TED while it is in the anion - SiF4 complex. As a
result, the addition of SiF4 is useful only for qualitative detection of low EA compounds.
Another means of minimizing the TED reaction of low EA compounds is to
simply lower the temperature of the ion source [80]. In this way, a point should be
reached where the lifetime of the molecular anion against TED would be long relative to
its lifetime against normal ion loss processes (either recombination with positive ions or
diffusion to the walls) and the mass spectra thereby produced should include an intense
molecular anion.
However, when using ion sources of relatively high pressure, as in APIMS or
IMS, along with relatively low ion source temperatures, we have found great difficulty in
observing molecular anions, as expected. For example, in spite of numerous attempts to
observe and study the molecular anions of aromatic compounds of EA lower than about
0.7 eV (such as benzophenone, anthracene and azulene), we have observed a molecular
anion only for the case of azulene [90], even with use of ion source temperatures as low
as 20°C. It is interesting to note that this observation was also made some 25 years ago by
Homing, Carrol and Dzidic [82], the first practitioners of APIMS.
In recent additional observations of our own by IMSZMS, we have also noted that
when low EA compounds, such as benzophenone and anthracene, are added to the ion
source at relatively low temperatures, the intensity of an ion at m/z = 50 is significantly
increased (Figures 3-1).
77
K inetic Ion M obility Spectra
Anthracene
>
300
-
200
-
Kinetic Ion M obility Spectra
Benzophenone
?
100
800
-
i l l Hii
mass (m/z)
Figure 3-1. IMS spectra of anthracene demonstrating the absence of an M m/z = 178
peak and the presence of the m/z = 50 peak. IMS spectra of benzophenone demonstrating
the absence of an M m/z = 182 peak and the presence of the m/z = 50 peak.
Since a likely assignment for the identity of this ion is (C^FbO)', this observation
suggests that if molecular anions are being produced in these cases, they are being rapidly
destroyed by reactions involving trace levels of oxygen and water both of which
commonly have partial pressures approaching the mTorr level in the buffer gases of one
atmosphere total pressure. In the present study, this possibility is explored in detail and is,
indeed, shown to explain the lack of REC responses to low EA compounds in high
pressure ion sources at low ion source temperatures. This study also reveals a novel
mechanism for the reaction of molecular anions with water and oxygen that is shown to
be uniquely fast for low EA compounds. It is also shown that the new reaction processes
revealed here offers a simple and reliable means for determining the EA of compounds
having EAs less than about 0.7 eV.
78
Experimental
A pulsed high pressure mass spectrometer (PHPMS) was used for all experiments.
The PHPMS was constructed in our laboratory and has been described in detail
previously [77, 91]. Chapter I describes how the instrument is set up as well as
summarizes the types of experiments and data that can be interrogated. For the present
experiments, a gaseous mixture consisting of small quantities of water, oxygen, and the
low EA compound of interest, M, were added to a methane buffer gas, in an associated
gas handling plant. This mixture then flowed slowly through the ion source of the
PHPMS. The ion source pressure was set to some constant value between I and 4 Torr.
The ion source temperature was generally held constant at 50°C. A brief pulse (20 (is) of
3000 eV electrons produced positive ions and electrons within the ion source. In the
methane buffer gas, these secondary electrons were rapidly thermalized and then captured
primarily by the compound M (Reaction 3-1) to form molecular anions, M", which are
also rapidly brought to thermal energy by collisions with the buffer gas. At the relatively
low temperature used, the primary loss of these M' ions will not be by their TED
reactions, but will be shown to be by reactions with oxygen and water. The number
density of ions within the source is sufficiently low so that the dominant loss of total
negative charge is by diffusion to the ion source walls rather than by recombination with
positive ions [8]. Relative ion abundances within the ion source are determined as a
function of time after the electron pulse by measuring the relative ion wall currents; that
is, by observing ,the ions that pass through a narrow slit on one wall of the ion source into
79
a vacuum chamber where the ions are mass analyzed (quadrupole mass filter), detected
(ion-counting channeltron) and time analyzed (multichannel scaler).
Results and Discussion
In order to mimic the absolute concentrations of water and oxygen that might
commonly be present in an ion source at one atmosphere pressure, small amounts of
water (0.9 mTorr) and oxygen (0.5 mTorr) were added to the methane buffer gas (3.0
Torr) of our PHPMS. The negative ion mass spectrum thereby produced is shown in
Figure 3-2. At the attenuation setting used for this mass spectrum, a very weak signal is
noted at m/z = 50, presumed to be due to the negative ion, (O20H2O)'.
This ion was possibly formed by the REC reaction of O2, followed by the
hydration of the resulting O2" ion by water. The observed low intensity of this ion can be
attributed to the very low rate constant for the REC reaction by oxygen. The pseudo
second-order rate constant for the attachment of thermal energy electrons to oxygen at
SO0C in nitrogen buffer gas at 3 Torr pressure is estimated to be only about I x 10"13 cm3
s"1 [92], which is about 6 orders of magnitude slower than the rate constants of fast REC
processes, for example, those of numerous substituted nitrobenzene compounds are
known to exceed I x IO"7 cm3 s’1[20].
The mass spectrum shown in Figure 3-2 B was then obtained after adding 0.06
mTorr of anthracene to the same gas mixture that was used to produce the spectrum in
Figure 3-2 A. Under these conditions, it is noted that only a small amount of the
molecular anion, M", for anthracene at m/z = 178 is detected.
80
3000 I
2500 -
2000
-
1500 -
1000
-
500 -
500 i
0
100
200
mass (m/z)
Figure 3-2. (A) The negative ion mass spectrum obtained with 0.5 mTorr oxygen and 0.9
mTorr water present in the methane buffer gas of the PHPMS ion source. The total
pressure is 3.0 Torr and the ion source temperature is 50°C. (B) The negative ion
spectrum obtained immediately after 0.06 mTorr anthracene was also added to the ion
source mixture described in part A.
The much more important effects of anthracene’s addition are noted for the ions of lower
mass, which are not materially related to anthracene. The most abundant of these is the
ion at m/z = 50, again thought to be (C^HaO)', and the next most abundant ion at m/z =
68 is thought to be (Oa^HaO^)’. An ion of low intensity at m/z = 60 is thought to be due
to CO3" which is probably also formed from (OaeHaO)" by its reaction with trace levels of
81
CO2 in the buffer gas. An ion of minor abundance at m/z = 126 is thought to be due to
presence of an unknown impurity possibly introduced with anthracene. Finally, an ion at
m/z = 210 is also invariably observed with anthracene and oxygen simultaneously present
in the ion source, although the intensity of this ions is very low. under the specific
conditions of the experiment shown in Figure 3-2 B. This ion is thought to be due to an
anthracene-oxygen adduct ion, (MoO2)". In summary, the spectra in Figure 3-2 are
roughly those which we have been typically observed whenever anthracene or other
compound of similarly low EA has been introduced to an atmospheric pressure ion
source set to a relatively low temperature. Instead of observing an intense M ' ion, as
expected, the intensity of the (0 2oH20)" ion is greatly increased.
Determination of the Mechanism
In order to obtain information concerning the dynamics of the reactions that
produced the unexpected spectrum shown in Figure 3-2, measurements of relative ion
intensities were also made as a function of time after the e-beam pulses. Again using the
case of anthracene as an example, one set of measurements of this type is shown in
Figure 3-3. Under the conditions of reagent concentrations used in this case, the
molecular anion of anthracene, M ", is the major ion initially formed by its REC reaction
immediately after the e-beam pulse. However, within only about 2 ms after the pulse, the
relative abundance of M" has been decreased to a terminal level of about 0.30 and that of
the (O2QH2O)" ion has increased to a terminal level of about 0.55.
82
time (msec)
Figure 3-3. PHPMS measurements of the relative intensities of the major ions, M " (x),
O2" (*), (M*02) " ( ) and (O2^H2O) (A), observed as a function of time after the e-beam
pulse with 0.80 mTorr oxygen, 0.050 mTorr water and 0.17 mTorr anthracene present in
the ion source. The total methane buffer gas pressure is 3.0 Torr and the ion source
temperature is 50°C.
Over this short period of time, an O2 ion and the anthracene-oxygen adduct ion, (M«02) ,
have also each reached terminal levels of about 0.07. Over the entire period from about 2
ms to 10 ms after the pulse, either a state of chemical equilibrium or a steady dynamic
state appears to have been reached in which the relative abundances of these four ions
remain constant.
In attempting to identify the detailed reactions and mechanism that produced the
spectrum in Figure 3-2 and the time dependencies shown in Figure 3-3, it is appropriate
to first consider the most obvious candidate which is primarily based on well-known
83
I
negative ion-molecule reactions. This mechanism will be referred to here as Model A and
consists of the series of reactions shown below:
M
+ O2 — M + O2
(3-2)
O 2 + H 2O
(3-3)
M" + O 2 ^ ( M o O 2)'
(3-4)
These reactions are assumed to occur immediately after the molecular anion, M', has
been formed by the e-beam pulse and Reaction 3-1. Reaction 3-2 is a simple electron
transfer from the molecular anion to oxygen. The rate constants for the forward and
reverse directions can be readily estimated. The reverse reaction is expected to occur with
collision frequency [93] and, therefore, will be about k_2 = 2 x IO 9cc s'1. The rate
constant for the forward reaction can be estimated from k_2 and the equilibrium constant,
K2 = k2 /k -2, expected for this reaction which can be deduced from Ki = exp (-AG°2 /
RT). The standard free energy change, AG°2, for Reaction 3-2 is adequately provided by
the difference in the electron affinities of oxygen and anthracene [73]. Using a literature
value of EAG2 = 0.45 eV [94, 95] and the EA of anthracene to be determined here, E A m
= 0.54 eV, an estimate of ki = 8 x IO"11cc s'1 at SO9C is obtained. Rate constants for both
the forward and reverse directions of Reaction 3-3 can be obtained from previous
determinations of the third-order rate constant for the forward direction [96] and the total
free energy change, AG0S= -11.9 kcal mol"1for this reaction at 50°C [97]. From this
information, a second order rate constant of about kg = 2 x 10"11 cc s"1 and a first order
84
rate constant of about k.3 = 4 s"1 are obtained under the present reaction conditions of 3.0
Torr total pressure and 50°C. The combination of Reaction 3-2 and 3-3 in Model A
provide for the conversion of M ' to (O2OH2O)". Reaction 3-4 is included in Model A in
order to account for the observed production of the adduct ion, (MoO2)", and as a side
reaction, is envisioned to play no role in the conversion of M "to (O2QH2O)" ions. In order
to account for the position of equilibrium for Reaction 3-4 and the rapid achievement of
this state observed in Figure 3-5, a near collisional second-order rate constant of Lt = I x
IO"9 cc s"1 and a first-order rate constant of L4 = 8 x IO4 s'1 have been assigned to the
forward and reverse directions of Reaction 3-4, respectively, in Model A. In order to
determine whether Model A provides an adequate explanation of the dynamics of the
anthracene reaction system, a computer simulation (using a deterministic differential
equation problem solving routine) of the pulsed e-beam experiment based on Model A
was created and is shown in Figure 3-4. In comparing this prediction with the
experimental results shown in Figure 3-3, it is noted that the final equilibrium state
predicted by Model A is, indeed, in good agreement with the experimental results.
However, it is also noted that the time required to achieve this terminal state by Model A
is about 200 times longer (0.4 s) than was required in the experiment (2 ms). Therefore,
Model A does not adequately describe the processes by which the M " ions of anthracene
were converted to (O2QH2O)" ion in Figure 3-2.
85
A nother candidate m echanism , which will be referred to here as M odel B, can be
envisioned sim ply by adding Reaction 3-5 to Reactions 3-2, 3-3 and 3-4 o f M odel A.
(M *02) + O2 ^ (O2-H 2O)
(3-5)
time (sec)
Figure 3-4. The relative intensities of the m ajor ions, M , O 2 ", (M*C>2 ) " and (O2^H2O)"
predicted by the candidate mechanism, M odel A, as a function o f tim e under the ion
source conditions described in Figure 3-3.
In the new model thereby created, Reaction 3-4, which is very fast in the forward
direction, along with Reaction 3-5 becomes the m ajor means for conversion of M to
(O 2eH2O)" ions. Reaction 3-5 might be viewed as a cluster exchange reaction in which a
86
water molecule replaces the molecule, M, in the MO2" complex ion. An interesting point
concerning Reaction 3-5 is that if the center of negative charge density within the
complex ion, (MoO2)", lies within the molecule M rather than O2 (as might be expected
from the fact that EAm > EAo2), then motion along the reaction coordinate towards the
transition state for Reaction 3-5 would involve a shift in negative charge density from the
M to O2 molecules within the complex ion. Alternately, it is also possible that the center
of negative charge within the (MoO2)" actually lies within the O2 species, in spite of its
lower EA, possibly due to the increased ion clustering potential of the molecule, M, over
that of O2.
In order to estimate rate constants for the forward and reverse directions of
Reaction 3-5, the same procedure as was previously applied to Reaction 3-4 will be used.
That is, it is assumed that the steady-state condition which is achieved by the
intermediate ion, (MoO2)", and the product ion, (O2OR2O)", in Figure 3-3 is an equilibrium
condition for Reaction 3-5. With this assumption, an equilibrium constant for Reaction 5,
K5 = k5/ k -5 = 25, is determined from the measured ratio of (MoO2)" and (O2OH2O)' ion
intensities in Figure 3-5 at any time after t = 2 ms and the known concentrations of H2O
and M within the ion source: The magnitudes of k5 and k_5were then varied until the
model provided the best fit to the observed time dependencies of ion intensities. With
estimates of k5 = 2 x IO"9 cc s’1 and k -5 = 8 x IO"11 cc s"1, and using the same rate
constants as used in Model A for Reactions 3-2, 3-3 and 3-4, the time dependencies
r
87
predicted by Model B are shown in Figure 3-5 to be in very good agreement with the
experimental results in Figure 3-3.
time (msec)
Figure 3-5. The relative intensities of the major ions, M ", O2", (M^Ch)' and (OaeH2O)"
predicted by the candidate mechanism. Model B, as a function of time under the ion
source conditions described in Figure 3-3.
In summary, while Reactions 2 and 3 must be occurring and account for the
steady-state abundance of O2 observed in this case, they are of minor importance in the
observed fast conversion of M "ions to O2 (H2O) ions. This is thought to be caused
primarily by the sequence of Reactions 4 and 5, both of which are sufficiently fast in both
directions as to cause the rapid achievement of a true equilibrium condition between
these two ions.
88
In Figure 3-6, a PHPMS experiment is shown where quinazoline rather than
anthracene was used as the low EA compound, relative intensity with the (MeO2)" ion
being significantly more intense than it was for anthracene in Figure 3-3. For this reason,
much less oxygen and much more water was used in order to have significant ion
abundances of these three major ions throughout the period of measurement.
’</>
C
C
0)
>
CQ
0)
time (msec)
Figure 3-6. PHPMS measurements of the relative intensities of the major ions, M ' (x),
(M*02) "(*), (M»02*H20) (-), (0 2*H20) (A), and (O2 ^(H2O)2)' (o), observed as a
function of time after the e-beam pulse with 0.15 mTorr oxygen, 1.06 mTorr water and
0.38 mTorr quinazoline present in the ion source. The total methane buffer gas pressure is
3.0 Torr and the ion source temperature is 50°C.
89
Under these conditions, the ions M ", (MoQ2)" and (O2OH2O)" again have significant
Because a higher water concentration was used in this case, two higher-order water
cluster ions, (MoO2OH2O)' and (O2O(H2O)2)", also have significant relative intensities in
Figure 3-6. The O2"ion does not have significant relative abundance because much less
oxygen was used and because the rate constant, k2, will be smaller than it was for
anthracene due to the slightly greater EA for quinazoline (0.56 eV). Figure 3-6 also
demonstrates that the time required for establishment of constant relative ion intensities is
very short, about 2 msec. Application of Model A to this reaction system predicts that
about 0.5 seconds would be required to achieve this terminal state and, therefore. Model
A again fails to explain this reaction system. However, when Reaction 5 is also included
in the mechanism for quinazoline, excellent agreement between the experimental results
and those predicted by Model B are obtained when values of ks = 2 x IO"10 cm3 s"1 and k.5
= 2 x IO"9 cm3 s"1 are assigned. In order to account for the two higher order water cluster
ions observed in this case, rate constants for the forward and reverse water clustering
reactions were set to 4.6 x IO"11 cm3 s"1 and 4.9 x IO3 s"1, respectively, values which have
been previously determined specifically for the clustering reaction leading to
(O2O(H2O)2)" at SO0C [96].
In Figure 3-7, a PHPMS experiment using benzophenone (BA = 0.61 eY) as the
Iow-EA compound is shown. In this case and under these experimental conditions, only
the M ", (MoO2)" and (0 2oH20)" ions have major relative abundance. Again, a steadystate condition is achieved, within about I msec after the e-beam pulse, and this
90
observation cannot be explained in terms of Model A which predicted about 0.7 seconds
to reach equilibrium.
2
V)
0)
0)
>
*3
CO
0)
time (m sec)
Figure 3-7. PHPMS measurements of the relative intensities of the major ions, M (x),
(M*02y (*) (O2WH2O)" (A), and (O2^(H2O)2) (o), observed as a function of time after the
e-beam pulse with 0.44 mTorr oxygen, 0.95 mTorr water and 0.86 mTorr benzophenone
present in the ion source. The total methane buffer gas pressure is 3.0 Torr and the ion
source temperature is 50°C.
However, the results in Figure 3-7 are also well explained by Model B when the
appropriate forward and reverse rate constants, shown later, are assigned to Reactions 4
and 5. In Figure 3-8, a PHPMS experiment using quinoxaline, which has a significantly
greater EA of 0.68 eV, is shown.
91
*
V)
C
0)
0)
>
5
5
time (msec)
Figure 3-8. PHPMS measurements of the relative intensities of the major ions, M (x),
(02*H20)" (A), and (02*(H20)2)‘ (o), observed as a function of time after the e-beam
pulse with 0.98 mTorr oxygen, 1.52 mTorr water and 0.040 mTorr quinoxaline present
in the ion source. The total methane buffer gas pressure is 3.0 Torr and the ion source
temperature is 50°C.
Even though relatively high concentrations of both oxygen and water were used, it is
noted that (C^eH2O) ions are formed somewhat slower than for they were with use of the
three compounds of lower EA previously considered. In this case, the achievement of
terminal ion abundance ratios required about 5 ms. Also, it is noted that the adduct ion,
(MeO2)", does not have significant relative abundance in this case. The fact that O2 is not
observed in this case is expected because quinoxaline’s higher EA shifts the equilibrium
92
position of Reaction 3-2 far to the left. As with the cases previously considered, the time.
dependence of the ion intensities observed for quinoxaline can not be explained by Model
A. Primarily because the magnitude of k2 for this reaction system is exceedingly low.
Model A predicts that about 0.6 seconds would be required to achieve the terminal state.
Therefore, it appears that this reaction system also proceeds by a different mechanism,
which might also be assumed to be Model B. In the case of quinoxaline, however, it was
not possible to experimentally determine the magnitudes of the individual rate constants
of the two elementary steps, Reaction 3-4 and 3-5, from PHPMS measurements as is was
in the previous cases considered because the intermediate adduct ion, (MoO2)', does not
have significantly high relative intensity as to provide a reliable estimate of the
equilibrium constants, K4 and K5.
In Figure 3-9, a PHPMS experiment using azulene as the low EA compound is
shown. Azulene also has a somewhat greater EA of 0.69 eV [98] and, in addition, is
known to have an usually large entropy of negative ionization. AS0n; = 4.5 cal K 1 mol"1
[98] which contributes about 0.08 eV additional free energy to its negative ionization at
SO0C (as described in the previous sections on thermal electron detachment and statistical
mechanics, this entropy term is not ASrxn, it is strictly the entropy difference of the neutral
molecule and its anion, or S0anIon - S0neUtrd = AS0in). These experimental conditions
93
1 .0 - r
time (msec)
Figure 3-9. PHPMS measurements of the relative intensities of the major ions, M (x),
(M*02)' (*) (O2^H2O) (A), and (O2^(H2O)2)' (o), observed as a function of time after the
e-beam pulse with 4.8 mTorr oxygen, 4.0 mTorr water and 0.040 mTorr azulene present
in the ion source. The total methane buffer gas pressure is 3.0 Torr and the ion source
temperature is 50°C.
included very high concentrations of both oxygen and water. As a result, (0 2*H20) ions
were formed much more slowly than for they were for all four compounds of lower EA
previously considered. After the 40-ms period of this experiment, a terminal steady-state
condition was being approached but has not yet clearly been achieved. It is also noted
that similar to the quinoxaline experiments, the adduct ion, (M»02)\ does not have
significant relative abundance. The fact that O2 is not observed is again expected due to
94
the higher EA and the A S ° ni of azulene, which shifts the equilibrium position of Reaction
3-2 far to the left. As with the other molecules previously considered, the time
dependence of the ion intensities observed for azulene could not be explained in terms of
Model A. Due primarily to the exceedingly small magnitude of k2 (about 3 x ICTi4 cm3
s"1) for this case, Model A predicts that the terminal state would take almost a full second
to reach. Therefore, Model B may be required to explain the reaction dynamics
associated with azulene. For this case, the rate constants for Reactions 3-4 and 3-5 again
could not be individually estimated because the intermediate ion, (MoO2)", did not have
significant relative intensity. The results shown here for azulene differ significantly from
all of those previously considered here in that even though very high oxygen and water
concentrations were used and the faster reaction dynamics of Model B were considered
operative, the conversion of M' ions to (O2oH20)" ions proceeded too slowly to bring
these reactions into equilibrium within the time of the experiment. These observations
explains why in past studies of the electron capture reactions of low EA compounds in
buffer gases of one atmospheric pressure containing trace levels of oxygen and water, the
molecular anion of azulene has been easy to observe while those of compounds having
slightly lower EA’s have not been.
Determination of Electron Affinities
It has been shown above that the molecular anions, M ", initially formed by
resonance electron capture reactions and the (0 2t>H20)" ions formed by subsequent
reactions of M " with oxygen and water can be brought into chemical equilibrium within
95
the time scale of the PHPMS experiments. The dominant processes by which this occurs
has been shown to be Reactions 3-4 and Reaction 3-5 (Model B). Therefore, an overall
expression for the equilibrium condition can be written by Combining Reactions 3-4 and
3-5 as shown by Reaction 3-6 below:
M + O2 + H 2 O
(C ^ 0H zjO ) + M
(3-6)
Equilibrium constants Ke (atm"1) for this overall process can then be obtained from the
PHPMS measurements and Equation 3-7:
Ke =
Ioa-(Hao) P m
/ I m - P0 2 Pmo
(3-7)
where I, is the relative intensity of the ion, i, after equilibrium has been achieved and P7-is
the partial pressure of the substance, 7, within the ion source in units of atmospheres. An
expression for the free energy change of Reaction 3-6, AG°6, can be written in terms of
the single step reactions in either candidate mechanism, Model A or Model B. Therefore,
A G 06 will
be given by:
A G 06 = A G 0Z + A G ° 3 .
For those cases except azulene where the
AS016 of M is of negligible magnitude, AG0Zwill be well-approximated by the difference
in electron affinities of oxygen and M; i.e., A G °z = EAm - EAoz- Since A G 's is known to
be -11.9 kcal mol"1 (-0.52 eV) at SO0C [97] and EAoz is known to be 0.45 eV [94, 95]
the electron affinity of the Iow-EA compound can be determined from the PHPMS
measurements of the equilibrium constant, K6, and Equation 3-8:
E A m — E A q 2 - AG 3 - RT In Kg
(3-8)
In Figure 3-10, the EAm values determined in this way are shown for all four of the low-
96
EA compounds for which a state of chemical equilibrium was achieved within the time
scale of the PHPMS experiments. For each compound, many determinations of this kind
0.70 -
0.65 -
0.60 -
0.55 -
Source Pressure (torr)
Figure 3-10. Electron affinity determinations for anthracene (□), benzophenone (o),
quinoxaline (A), and quinazoline (x) by the PHPMS method described here. The
individual measurements shown where obtained at 50° C using a variety of different
reagent concentrations and total ion source pressures between I and 4 Torr.
were made using different combinations of concentrations for oxygen, water and the low
EA compound. In addition, the total ion source pressure was also varied over the range
from I to 4 Torr. As can be seen in Figure 3-10, the EA values thereby determined where
relatively independent of these changes in experimental conditions.
97
EA values were determined for each molecule from the average of these
measurements. These values of 0.54, 0.56, 0.61 and 0.68 eV for anthracene, quinazoline,
benzophenone and quinoxahne, respectively^ are shown to be in reasonably good
agreement of the literature values (anthracene [99, 100],quinazohne [101], benzophenone
[93, 94, 102],quinoxaline [101]).
The method of EA determination for Iow-EA compounds just described offers an
advantage over the usual PHPMS method for EA determinations in which the equilibrium
position of the electron transfer reaction of the compound of interest is measured against
another compound of known and similar EA[93], In such measurements, it is often
difficult to arrange the concentration ratios so that significant ion intensities are observed
for both molecular anions (as in the case of Reaction 3-2 in the present study). With the
present method and the simultaneous use of two reference compounds, oxygen and water,
a given change of the concentrations of both oxygen and water results in a much greater
change in the ion intensity ratio, Ioz-(Hzo) / I m - (for example, lowering the concentrations of
O2 and H2O by one order of magnitude causes an increase in the ion intensity ratio of two
orders of magnitude). In addition, with use of the conventional method based on paired
low EA compounds, the ubiquitous presence of trace oxygen and water in common buffer
gas supplies give rise to the fast reactions described here which, of course, constitute
unwanted side reactions in the conventional method.
98
Conclusions
In this study, we have discovered why molecular anions, M ", for many low EA
compounds, M, are not readily observed in electron capture ion sources operating under
buffer gas conditions of near ambient conditions of pressure and temperature. This is
because these molecular anions react rapidly with trace levels of oxygen and water to
form the ion, (Oz0HhO)". It has been shown that the conversion of M ' to (Oz0HzO)' ions
proceeds by way of a two-step mechanism in which the an intermediate ion complex of
the type, (MoOz)", is first formed by the reaction of M ' with oxygen. Because forward
and reverse rate constants for both of the elementary steps in this mechanism are
relatively large, a state of chemical equilibrium is readily achieved for the overall process
and this provides a convenient means for determining the electron affinities of the low
EA compounds. By this method, the electron affinities of anthracene, quinazoline,
benzophenone and quinoxaline are shown here to be equal to 0.54, 0.56, 0.61 and 0.68
eV, respectively.
99
REFERENCES CITED
100
REFERENCES CITED
l.Wolpert, L. and A. Richards, Passionate minds : the inner world of scientists. 1997,
Oxford ; New York: Oxford University Press, vi, 240 p.
2. Homing, E.C., et al, Electron capture : theory and practice in chromatography.
Journal of chromatography library, ed. A. Zlatkis, Poole, C. F. Vol. 20. 1981,
Amsterdam ; New York: Elsevier Scientific Pub. Co. xii, 429.
3. Carson, R., Silent spring. 40th anniversary ed. 2002, Boston: Houghton Mifflin, xix,
378 p.
4. Christophorou, L.G., Electron-molecule interactions and their applications. 1984,
New York: Academic Press.
5. Gas Lasers, in Applied atomic collision physics Pure and applied physics; v. 43,
H.S.W. Massey, E.W. McDaniel, and B. Bederson, Editors. 1982, Academic
Press: New York.
6. Manos, D M. and D.L. Flamm, Plasma etching : an introduction. Plasma —materials
interactions. 1989, Boston: Academic Press, xii, 476 p.
7. Wayne, R.P., Chemistry of atmospheres : an introduction to the chemistry of the
atmospheres of earth, the planets, and their satellites. 3rd ed. 2000, Oxford
[England] ; New York: Oxford University Press, xxix, 775 p.
8. Kebarle, P., Pulsed High Pressure Mass Spectrometer, in Techniques for the Study of
Ion Molecule Reactions, J.F.aW.S. Jr., Editor. 1989, CRC Press: Boca Raton, p.
221-286.
9. Williamson, D.H., W.B. Knighton, and E.P. Grimsrud, Effect of buffer gas alterations
on the thermal electron attachment and detachment reactions of azulene by
pulsed high pressure mass spectrometry. International Journal of Mass
Spectrometry, 2000. 196: p. 481-489.
101
10. Munson, M.S. and F.H. Field, Chemical Ionization Mass Spectrometry. Journal of the
American Chemical Society, 1966. 88(12): p. 2621-30.
11. McDaniel, E.W. and E.A. Mason, The mobility and diffusion of ions in gases. 1973,
New York,: Wiley, xi, 372 p.
12. Mason, E.A., Ion mobility: its role in plasma chromatography, in Plasma
chromatography, T.W. Carr, Editor. 1984, Plenum Press: New York. p. 43-93.
13. Kregel, M.D., Diffusion in Decaying Plasmas. Journal of Applied Physics, 1970.
41(5): p. 1978-&.
14. McDaniel, E.W., Collision phenomena in ionized gases. Wiley series in plasma
physics. 1964, New York,: Wiley, xxii, 775,p.
15. Fanklin, R.N., Electronegative plasmas —why are they so different? Plasma Sources
Science & Technology, 2002. 11(A): p. A31-A37.
16. Biondi, M.A., Dissociate attachment of electrons in iodine. I. Microwave
determination of the absolute cross section at 300K. Physical Review, 1959.
109(6): p. 2005.
17. Miller, T.M., et at, Observation of Thermal Electron Detachment from CycZo-QFgin Falp Experiments. International Journal of Mass Spectrometry and Ion
Processes, 1994. 135(2-3): p. 195-205.
18. Rogoff, G E., Ambipolar Diffusion-Coefficientsfor Discharges in Attaching Gases.
Journal of Physics D-Applied Physics, 1985. 18(8): p. 1533-1545.
19. Lineberger, W.C. and L.J. Puckett, Positive ions in nitric oxide afterglows. Physical
Review, 1969. 186(1): p. 116.
20. Knighton, W.B., et al., Resonance Electron-Capture Rate Constants for Substituted
Nitrob'enzenes. Journal of Physical Chemistry, 1994. 98(14): p. 3770-3776.
102
21. Vogt, E., Wannier, G.H., Physical Review, 1954. 95: p. 1190.
22. Dunning, P B., Electron-Molecule Collisions at Very-Low Electron Energies. Journal
of Physics B-Atomic Molecular and Optical Physics, 1995. 28(9): p. 1645-1672.
23. Klots, CE., Rate constants for unimolecular decomposition at threshhold. Chemical
Physics Letters, 1976. 38(1): p. 61.
24. Compton, R.N., et al., Nondissociative electron capture in complex molecules and
negative ion lifetimes. The Journal of Chemical Physics, 1966. 45(12): p. 4634.
25. Hahndorf, I. and E. Hlenberger, Temperature dependence of electron attachment
processes. International Journal of Mass Spectrometry, 1997. 167/168: p. 87.
26. Christophorou, L.G. and J.K. Olthoff, Electron interactions with SFg. Journal of
Physical Chemistry Reference Data, 2000. 29(3): p. 267.
27. Takayanagi, K., Introduction to electron-molecule collisions, in Electron-Molecule
Collisions, K. Takayanagi and I. Shimamura, Editors. 1984, Plenum: New York.
p. 1-80.
28. Barsotti, S., M.W. Ruf, and H. Hotop, Clear experimental evidence for p-wave
attachment- Threshhold behavior in electron attachment to chlorine molecules.
Physical Review Letters, 2002. 89(8): p. 08201-1.
29. Reddish, T., Electron and photoelectron spectroscopy, in Unpublished Class Notes
Windsor University.
30. Thoss, M. and W. Domcke, Theory of vibrational relaxation processes in resonant
collisions of low energy electrons with large molecules. Journal of Chemical
Physics, 1998. 109(16): p. 6577.
31. Dessent, C.E.H., J. Kim, and M.A. Johnson, Spectroscopic observation of vibrational
Feshbach resonances in near-threshold photoexcitation of X dot CH3NO2 (X- = T
and Br). Faraday Discussions, 2000(115): p. 395-406.
103
32. Leber, E., et al., Vibrational Feshbach resonances in electron attachment to nitrous
oxide clusters: decay inter heterogenous and homogeneous cluster anions.
Chemical Physics Letters, 2000. 325: p. 345.
33. Leber, E., et al., Vibrational Feshbach resonances in electron attachment to carbon
dioxide clusters. European Physical Journal D, 2000. 12(1): p. 125-131.
34. Tulej, M., et al., Feshbach states of the propadienylidene anion H2 CCC. Physical
Chemistry Chemical Physics, 2001. 3: p. 4674.
35. Miller, T.M., et al., Electron attachment to SF5 CF3 (296-563 K) and calculations of
the neutral and anion thermochemistry. Journal of Chemical Physics, 2002.
116(14): p. 6021-6027.
36. Miller, T.M., et al., Electron attachment to PC13 andPOCl3, 296-552 K. Journal of
Chemical Physics, 1998. 109(2): p. 578-584.
37. Miller, T.M., et al., Thermal Electron-Attachment to SF4 and SFg. Journal of
Chemical Physics, 1994. 100(12): p. 8841-8848.
38. Suess, L., R. Parathasarathy, and F.B. Dunning, Nondissociative low energy electron
attachment to SF^,
CjoFg, and C-CyFjp Negative ion lifetimes. Journal of
Chemical Physics, 2002. 117(24): p. 11222.
39. Michaud, M., M. Lepage, and L. Sanche, Lifetime of negative ion resonances and the
density of free electron states: Oy isolated in an argon matrix. Physical Review
Letters, 1998. 81(13): p. 2807.
40. Naff, W.T. and C.D. Cooper, Transient negative-ion states in alicyclic and aromatic
fluorocarbon molecules. The Journal of Chemical Physics, 1968. 49(6): p. 2784.
41. Klots, C.E., Statistical aspects of autoinization lifetimes. The Journal of Chemical
Physics, 1967. 46(3): p. 1197.
104
42. Hadijiantoniou, A., L.G. Chnstophorou, and J.G. Carter, Long-lived parent negative
ions formed via nuclear-excited Feshbach resonances. Part 1-Benzene
derivatives. Journal of the Chemical Society. Faraday Transactions II, 1973. 69: p.
1691.
43. Hadijiantoniou, A., L.G. Christophorou, and J.G. Carter, Long-lived parent negative
ions formed via nuclear-excited Feshbach resonances. Part 2-Aromatic molecules
other than benzene derivatives and non-aromatic organic structures. Journal of
the Chemical Society. Faraday Transactions II, 1973. 69: p. 1704.
44. Collins, P.M., et al., Energy dependence of the electron attachment cross section and
the transient negative ion lifetimefor -benzoquinone and 1,4 -naphthoquinone.
Chemical Physical Letters, 1970. 4(10): p. 646.
45 Johnson, J.P., et al., Long-lived parent negative ions formed via nuclear-excited
Feshbach resonances. Part 4- Sytematic study of .N0 2 -containing benzene
derivatives. Journal of the Chemical Society. Faraday Transactions II, 1975: P1742.
46. Horacek, J., W. Domcke, and H. Nakamura, Electron attachment and vibrational
excitation in hydrogen iodide: calculations based on the nonlocal resonance
model. Zeitschrift, Fur Physik D, 1997. 42: p. 181.
47. Garrett, W.R., Thermally stable negative ions of polar molecules. Journal of
Chemical Physics, 1978. 69(6): p. 2621.
48. Tobita, S., et al., Polycyclic aromatic hydrocarbons: negative ion formation following
low energy (0-15 eV) electron impact. Chemical Physics, 1992. 161: p. 501.
49. Christophorou, L.G., A. Hadijiantoniou, and J.G. Carter, Long-lived parent negative
ions formed via nuclear-excited Feshbach resonances. Part 3-Variation of the
autodetachment lifetime with incident electron energy. Journal of the Chemical
Society. Faraday Transactions II, 1973. 69: p. 1713.
50. Atkins, P.W. and J. De Pmla., Atkins' Physical chemistry. 7th ed. 2002, Oxford ; New
York: Oxford University Press, xxi, 1149 p.
105
51. Lindermann, F.A., et al., Discussion on "The radiation theory of chemical action".
Transactions of the Faraday Soceity, 1922. 17: p. 598.
52. Slater, N.B., Theory of unimolecular reactions. 1959, Ithaca, N.Y.,: Cornell
University Press. 230 p.
53. Cleve, P.T. Les Prix Nobel. 1992. Stockholm,: Imprimerie Royale.
54. Miller, T M., Detachment —Personal Communication with Knighton, W.B. 1999:
Boston, MA.
55. Grimsrud, E.P., S. Chowdhury, and P. Kebarle, Thermal energy electron detachment
rate constants. The electron detachment from azulene' and the electron affinity of
azulene. Journal of Chemical Physics, 1985. 83(8): p. 3983-3989,
56. Van Doren, J.M., et al., ElectronAttachment to c yclo ^ F ^ h - International Journal
of Mass Spectrometry, 2003. Accepted but not yet published.
57. McQuarrie, D.A., Statistical thermodynamics. 1973, Mill Valley, Calif.: University
Science Books, xi, 343 p.
58. Knighton, W.B., J.A. Bognar, and E.P. Grimsrud, Thermal electron detachment rate
constants of the molecular anion of perfluorobenzene. Chemical Physics Letters,
1992.192(5,6): p.522.
59. Mock, R.S. and E.P. Grimsrud, Thermal Electron Detachment Rate Constants for
Azulene- at Atmospheric-Pressure by the Photodetachment-Modulated ElectronCapture Detector. International Journal of Mass Spectrometry and Ion Processes,
1989. 94(3): p. 293-303.
60. Kestin, J. and J.R. Dorfman, A course in statistical thermodynamics. 1971, New
York,: Academic Press, xv, 577 p.
61. Dow Chemical Company. Thermal Research Laboratory., JANAF thermochemical
tables. 2d ed. 1971, Washington,: U.S. National Bureau of Standards; for sale by
the Supt. of Docs., U.S. Govt. Print. Off. I v. (unpaged).
106
62. Chutjian, A., A. Garscadden, and J.M. Wadehra5ElectronAttachments to molecules
at low elctron energies. Physics Reports, 1996. 264: p. 393.
63. Fehsenfeld5F.C., Electron Attachment to SF<s. Journal of Chemical Physics, 1970.
53(5): p. 2000.
64. Alge5E., N.G. Adams5and D. Smith5Rate coefficients for the attachment reactions of
electrons with c-Cjf1 4 , CHjBr, CFjBr, CHjBrj, and CHjl determined between
200K and 600K using the FALP techniques. Journal of Physics B - Atomic
Molecular and Optical Physics, 1984. 17(18): p. 3827.
65. Alge5E., N.G. Adams5and D. Smith5Rate coefficients for the attachment reactions of
electrons with CCU, CCljF, CCljFj, CHClj, Clj and SFa determined between
200K and 600K using the FALP techniques. Journal of Physics B - Atomic
Molecular and Optical Physics, 1984. 17: p. 461.
66. Spanel5P. and D. Smith5Falp Studies of Electron-Attachment at Elevated Electron
Temperatures - the Influence of Attachment on Electron-Energy Distributions.
International Journal of Mass Spectrometry and Ion Processes, 1993. 129: p. 193203.
67. Smith5D. and P. Spanel5Studies of Electron-Attachment at Thermal Energies Using
the Flowing Afterglow Langmuir Probe Technique. Advances in Atomic,
Molecular, and Optical Physics, Vol 32, 1994. 32: p. 307-343.
68. Spanel5P., et al., A coordinatedflowing afterglow and crossed beam study of electron
attachment to CCBBr. International Journal of Mass Spectrometry, 1997. 167: p.
1- 12.
69. Hanstorp5D. and M. Gustafsson5Determination of the Electron Affinity of Iodine.
Journal of Physics B51992: p. 1773.
70. Christophorou, L.G. and J.K. Olthoff5Electron attachment cross sections and
negative ion states of S E International Journal of Mass Spectrometry, 2001.
205(1-3): p. 27-41.
107
71. Smith, D., et al., Formation of SF5 - in Electron-Attachment to SFg - Swarm and Beam
Results Reconciled. Chemical Physics Letters, 1995. 240(5-6): p. 481-488.
72. Culbertson, J.A. and E.P. Grimsrud, Fates of the Molecular Anions of SFg and CjF1 4
Upon Recombination with Positive-Ions. International Journal of Mass
Spectrometry and Ion Processes, 1995. 150: p. 87-98.
73. Kebarle, P. and S. Chowdhury, Electron Affinities and Electron Transfer Reactions.
Chemical Reviews, 1987. 87: p. 513-534.
74. Grimsrud, E.P., S. Chowdhury, and P. Kebarle, Electron-Affinity of SFg and
Perfluoromethylcyclohexane - the Unusual Kinetics of Electron-Transfer
Reactions a-+B=a+B-, Where a=SFg or Perfluorinated Cydo-Alkanes. Journal of
Chemical Physics, 1985. 83(3): p. 1059-1068.
75. Neumaier, M., O. Hampe, and M.M. Kappes, Electron Transfer Collisions of
Fullerene Dianions and SFg in a Penning-Trap-Personal Communication with
Salyards, M. J. 2003: Bozeman, MT.
76. Salyards, M.J., W.B. Knighton, and E.P. Grimsrud, Effects of oxygen and water on
the resonance electron capture reactions of low electron affinity compounds.
International Journal of Mass Spectrometry, 2003. 222(1-3): p. 201-212.
77. Williamson, D.H., W.B. Knighton, and E.P. Grimsrud, Lewis acid-base interactions
ofSip 4 with molecular anions formed by electron capture reactions. International
Journal of Mass Spectrometry, 2001. 206(1-2): p. 53-61.
78. Stemmier, E.A. and RA. Hites, Electron capture negative ion mass spectra of
environmental contaminants and related compounds. 1988, New York: VCH
Publishers, xvii, 390 p.
79. Sears, L.J., J.A. Campbell, and E.P. Grimsrud, Ionization Dynamics within the HighPressure Electron-Capture Mass-Spectrometer - the Unusual Spectra of
Derivatized Polycyclic Aromatic-Amines and Perchlorinated UnsaturatedHydrocarbons. Biomedical and Environmental Mass Spectrometry, 1987. 14(8):
p. 401-415.
108
80. Knighton, W.B., L J. Sears, and E.P. Grimsrud, High-pressure electron capture mass
spectrometry. Mass Spectrometry Reviews, 1995. 14(4-5): p. 327-343.
81. Dougherty, R.C., I. Dalton, and F.J. Biros, Negative Chemical Ionization MassSpectra of Polycyclic Chlorinated Insecticides. Organic Mass Spectrometry,
1972. 6(11): p. 1171.
82. E.C. Homing, D.I.C., I. Dzidic, RN. Stillwell, Electron capture : theory and practice
in chromatography, in Journal of chromatography library ; v. 20, C F. Poole,
Editor. 1981, Elsevier Scientific Pub. Co.: Amsterdam ; New York. p. xii, 429 p.
83. Dressier, M., Selective gas chromatographic detectors. Journal of chromatography
library. Vol. 36. 1986, Amsterdam; New York New York, NY, U.S.A.: Elsevier;
Distributors for the U.S. and Canada, Elsevier Science, xiii, 319.
84. Grimsrud, E.P., Detectors for capillary chromatography, in Chemical analysis; v.
121, H.H. Hill and D.G. McMinn, Editors. 1992, Wiley: New York. p. xvii, 444 p.
85. Eiceman, GA. and Z. Karpas, Ion mobility spectrometry. 1994, Boca Raton: CRC
Press. 228 p.
86. Bruins, A.P., Mass-Spectrometry with Ion Sources Operating at AtmosphericPressure. Mass Spectrometry Reviews, 1991. 10(1): p. 53-77.
87. Giles, K. and E.P. Grimsrud, The kinetic ion mobility mass-spectrometer Measurements of ion molecule reaction rate constants at atmospheric pressure.
Journal of Physical Chemistry, 1992. 96(16): p. 6680-6687.
88. Knighton, W.B., J.A. Bognar, and E.P. Grimsrud, Reactions of Selected Molecular
Anions with Oxygen. Journal of Mass Spectrometry, 1995. 30(4): p. 557-562.
89. Knighton, W.B., D.R. Zook, and E.P. Grimsrud, Cluster-assisted decomposition
reaction of the molecular anions of SF^ and CyF1 4 . Journal of the American
Society for Mass Spectrometry, 1990. 1(5): p. 372-381.
109
90. Sahlstrom, K.E., W.B. Knighton, and E.P. Grimsmd, Thermal electron detachment of
the molecular anion of azulene at elevated pressures by ion mobility
spectrometry. International Journal of Mass Spectrometry,. 1998. 180: p. 117-127.
91. Knighton, W.B. and E.P. Grimsmd, Gas-phase electron-transfer reactions between
selected molecular anions and halogenated methanes. Journal of the American
Chemical Society, 1992. vll4(n7): p. p2336(7).
92. Chanin, L.M., Phelps, A.V., Biondi, M B., Measurement of resonance electron
capture rates constants for oxygen. Physical Review, 1962. 128: p. 219.
93. Grimsmd, E.P., et al., Electron-Affinities from Electron-Transfer Equilibria - a+B=a+B-. Journal of the American Chemical.Society, 1985. 107(16): p. 46274634.
94. Chen, E.C.M. and W.E. Wentworth, Determination of Molecular Electron-Affinities
Using the Electron-Capture Detector in the Pulse Sampling Mode at Steady-State.
Journal of Physical Chemistry, 1983. 87(1): p. 45-49.
95. Travers, M.J., D.C. Cowles, and G.B. Ellison, Reinvestigation of the electron
affinities of O2 an NO. Chemical Physics Letters, 1989. 165(5): p. 449.
96. Payzant, J.D. and P. Kebarle, Kinetics of Reactions leading to Of(HiO) In Moist
Oxygen. Journal of Chemical Physics, 1972. 56((7)): p. 3482.
97. Arsahdi, M. and P. Kebarle, Hydration of OH and O f in gas phase - Comparative
solvation of OH by water and hydrogen halides- Effects of acidity. Journal of
Physical Chemistry, 1970. 74(7): p. 1483.
98. Chowdhury, S., et al., Entropy Changes and Electron-Affinities from Gas-Phase
Electron-Transfer Equilibria - a-+B=a+B-. Journal of Physical Chemistry, 1986.
90(12): p. 2747-2752.
99. Heins, T., Chowdhury, S., Kebarle P., Electron-affinities of napthalene, anthracene
and substituted napthalenes and anthracenes. Organic Mass Spectrometry, 1993.
28(4): p. 358-365.
no
100. Schiedt, I. and R. Weinkauf, Photodetachment phoioelectron spectroscopy of mass
selected anions: Anthracene and the anthracene-H20 cluster. Chemical Physics
Letters, 1997. 266(1-2): p. 201-205.
101. Dillow, G.W. and P. Kebarle, Electron-affinities of aza-substituted polycyclic
aromatic-hydrocarbons. Canadian Journal Of Chemistry-Revue Canadienne De
Chimie, 1989. 67(10): p. 1628-1631.
102. Huh, C., et al., Thermodynamic stabilities and resonance demand of aromatic
radical anions in the gas phase. Bulletin Of The Chemical Society Of Japan,
1999. 72(5): p. 1083-1091.
APPENDICES
A P P E N D IX A
HIGH PRESSURE LIMIT EQUATIONS
F rom R ea ction 1-7:
ki
k2[B]
M + e' ^ M*" —M:
k-1
k .2[B]
at
at
Steady state assumption:
^
u
/Ci[M ][g -] + fc.2[j?j[M - ]
k _ , + k 2l B ]
Substitute in [M*"]
J[M] =
dt
+
k_x + ^ 2[B]
Common denominator and related algebra follows.
J [M ]
k J ^ M ^ e
] + k _ , k _ 2 [ B ] [ M - ] - k _ lk l [ M ] [ e - ] - k 2 m k l [ M - \ [ e - ' \
/c_| +
dt
[5 ]
^[M] _ Al1Al2[B][M~]-A:2
^
+ ^2[S]
[Af ][e~]
Let the pressure of B be infinitely high for the high pressure limit
”
A"_, +
A2
So, in the high pressure limit:
at
= -^ [M ][e-] + L 1^ [ M - ]
Jcn
From the simplest Reaction 1-1:
kec
M + e' ^ M'
kdet
115
where k%is a function of a cap , the electron capture cross section, and
k
is a Maxwellk2
Boltzman function of temperature. It does not matter if the initial rate equation is written
in terms of [M"] instead [M]; the final relationships are unchanged.
A P P E N D IX B
REVIEW OF PSUEDO 1st ORDER KINETICS
117
d [ BP~ ]
= K m [SF6UBP- ]
dt
Km ISF6J = k
I a h dlip ^ k r s p j d '
A ln[ BP~ ]
At
ktrans
k
[SF6J
A P P E N D IX C
MODELING WITH DIFFERENTIAL EQUATIONS
119
= KcW ] [e- j - ^1J M - Jd[e
]
K a [ M ~ } ~ k ec[ M
dt
] [e _] -
km
[A T ]- kco,
K diffI e - ] - k ^ S F ^ e K
where kec is the electron capture rate constant for M, Icdet is the thermal detachment rate
constant, k## is the diffusion rate constant, kCoi is the Langevan collisional rate constant ~
2.0 x IO"9 cc sec"1, X is the concentration of any electron transferring impurities, IcedIff is
the elctron diffusion rate constant, and ka is the electron capture rate constant for SFd.
We assume that [e"J is steady state so
[eK
d [e-\
dt
KclM] + K<uff+KlSF6I
Substitute in for [e"]:
d[M~]
dt
C
=Kxm
Collect terms:
k ec[ M ] + k ediff + k a [ S F6\ ^
0 and can solve for [e‘]
120
d[M -\
r
\
= [M"]
~ k det - k diff - k col{ X ]
V
k ec I M
]
+
k ediff
+
k a[SF 5 ]
J
Separate variables and integrate:
U r ™ ''4 K c[ M ] + kediff + k a[ SF6].
0
f
ln[ M J t =
k/iet ~ kcliff ~ Koi [ X ] d t
ec
KXJMl
k J M ] + km + kJ SF6]
k da ~ k<njf - J i I x I
t + ln[ M
Assume electron diffusion is not happening and no impurities are present (i.e. kediff = 0,
and [X] = 0).
AlnfM-J
At
K etK J M ]
+ ki.j* =
diff
- k ,
K J M ] + Ica[ S F J
Common denominator and related algebra:
Mn] M J
At
_
diff
- K eJ J S F J
K J M ] + KJ SFJ
'^det
/„
Invert both sides:
r Aln[M~]
At
k J M ] + ka[SF6]
^
J
tK
[ S F 6 ]
Rearrange by dividing each term on the right side by Icec:
f Aln[M~] , V l
At
+kdiffj
[ M ] + ^ [ S F 6]
Kc
If
If
- jk L ^ [ SF6 ]
Rearrange into a linear equation:
C
V
,
r ,Ar- i
Mnl
M- J ,
a
A,
,
+ k‘«
A "1
f
K e A
J
V
ee
\r [M]
y
\
[SF6]
6J J
\ L
x +b
I
+
"det
Download